[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill-nvidia-vss-deploy-dense-captioning":3,"mdc-dy2ih6-key":34,"related-repo-nvidia-vss-deploy-dense-captioning":3011,"related-org-nvidia-vss-deploy-dense-captioning":3113},{"slug":4,"name":4,"fn":5,"description":6,"org":7,"tags":11,"stars":23,"repoUrl":24,"updatedAt":25,"license":26,"forks":27,"topics":28,"repo":29,"sourceUrl":32,"mdContent":33},"vss-deploy-dense-captioning","deploy RT-VLM dense captioning services","Use this skill when deploying standalone RT-VLM dense captioning or calling its REST API (uploads, captions, streams, chat-completions, Kafka). Not for VSS profile deploy or video-search ingestion.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},"nvidia","NVIDIA","https:\u002F\u002Fpexgzepcugksgbtrxkhf.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Forg-logos\u002Fnvidia.png",[12,16,19,20],{"name":13,"slug":14,"type":15},"Deployment","deployment","tag",{"name":17,"slug":18,"type":15},"REST API","rest-api",{"name":9,"slug":8,"type":15},{"name":21,"slug":22,"type":15},"Computer Vision","computer-vision",2473,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fskills","2026-07-17T05:29:01.916878","Apache-2.0",281,[],{"repoUrl":24,"stars":23,"forks":27,"topics":30,"description":31},[],"AI agent skills published by NVIDIA","https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fskills\u002Ftree\u002FHEAD\u002Fskills\u002Fvss-deploy-dense-captioning","---\nname: vss-deploy-dense-captioning\ndescription: Use this skill when deploying standalone RT-VLM dense captioning or calling its REST API (uploads, captions, streams, chat-completions, Kafka). Not for VSS profile deploy or video-search ingestion.\nlicense: Apache-2.0\nmetadata:\n  version: \"3.2.1\"\n  github-url: \"https:\u002F\u002Fgithub.com\u002FNVIDIA-AI-Blueprints\u002Fvideo-search-and-summarization\"\n  tags: \"nvidia blueprint operational deployment\"\n---\n## Purpose\n\nStand up the RT-VLM dense-captioning microservice on its own and exercise every endpoint it exposes (file upload, generate_captions, stream add\u002Fdelete, chat-completions, Kafka topics).\n\n## Prerequisites\n\nFor standalone RT-VLM deployment:\n- Docker, Docker Compose, NVIDIA Container Toolkit, and a visible GPU.\n- NGC registry credentials in `$NGC_CLI_API_KEY` for `docker login nvcr.io`,\n  image pulls, and local NGC model\u002Fartifact downloads.\n- `curl`, `jq`, and any writable working directory for the standalone compose copy.\n\nFor API calls against an existing service:\n- Running RT-VLM service reachable at `$BASE_URL`.\n- Bearer token in `$RTVI_VLM_API_KEY` or `$NGC_CLI_API_KEY`, depending on how the\n  service was configured.\n\nFor full VSS profile deployment:\n- Use `..\u002Fvss-deploy-profile\u002FSKILL.md`; this skill does not deploy full VSS profiles.\n\n## Instructions\n\nFollow the routing tables and step-by-step workflows below. Each section that ends in *workflow*, *quick start*, or *flow* is intended to be executed top-to-bottom. Detailed reference material lives in `references\u002F`; execute the documented workflows directly unless a future revision names a concrete helper.\n\n## Examples\n\nWorked end-to-end examples are kept under `evals\u002F` (each `*.json` manifest contains a runnable scenario) and inline in the per-workflow `curl` blocks below. Run a Tier-3 evaluation with `nv-base validate \u003Cthis-skill-dir> --agent-eval` to replay them.\n\n## Limitations\n\n- Requires either a standalone RT-VLM service deployed via this skill or an\n  existing RT-VLM service reachable from the caller.\n- NGC-hosted models and NIMs may be subject to rate-limits, GPU memory requirements, and license restrictions.\n- Concurrency, GPU memory, and storage limits depend on the host hardware and the profile's compose file.\n- Keep `NGC_CLI_API_KEY`, `RTVI_VLM_API_KEY`, and `rtvi-vlm.env` files out of git and out of logs; do not echo credential values or include them in final responses.\n- Docker group access and `sudo` are effectively root-level privileges. Use the non-interactive `sudo -n` guard in the deploy reference and stop for host-owner action when passwordless sudo is unavailable.\n\n## Troubleshooting\n\n- **Error**: REST call returns connection refused. **Cause**: target microservice not running. **Solution**: probe `\u002Fdocs` or `\u002Fhealth`; redeploy via `vss-deploy-profile` or the matching `vss-deploy-*` skill.\n- **Error**: HTTP 401\u002F403 from NGC pulls. **Cause**: missing\u002Fexpired `NGC_CLI_API_KEY`. **Solution**: `docker login nvcr.io` and re-export the key before retrying.\n- **Error**: container OOM or model fails to load. **Cause**: insufficient GPU memory for the selected profile. **Solution**: switch to a smaller variant or free GPUs via `docker compose down`.\n\n# Deploy and Use RT-VLM Dense Captioning (VSS 3.2)\n\nRT-VLM is NVIDIA's real-time vision-language microservice: decode video (file or\nRTSP), segment it into chunks, run a VLM (`cosmos-reason1`, `cosmos-reason2`, `cosmos-reason3`, or any\nOpenAI-compatible model), stream dense captions back over SSE\u002FHTTP, and publish\ncaptions, incident alerts, and errors to Kafka. Use this skill to deploy the\nstandalone RT-VLM service when a full VSS profile is not already running, then call\nits `\u002Fv1\u002F...` API for caption generation, file upload, live-stream management, health\nchecks, NIM-compatible chat completions, or Prometheus metrics. API reference:\n\u003Chttps:\u002F\u002Fdocs.nvidia.com\u002Fvss\u002Flatest\u002Freal-time-vlm-api.html>.\n\n## Deployment Routing\n\nIf the user asks to deploy a full VSS profile, use\n[`..\u002Fvss-deploy-profile\u002FSKILL.md`](..\u002Fvss-deploy-profile\u002FSKILL.md). That skill\nowns profile routing, `generated.env`, `resolved.yml`, multi-service sizing, and\nfull-stack deploy\u002Fteardown.\n\nIf the user asks for standalone RT-VLM dense captioning, or no VSS profile is\nalready running, use the standalone RT-VLM flow in\n[`references\u002Fdeploy-rt-vlm-service.md`](references\u002Fdeploy-rt-vlm-service.md)\nbefore calling the API. This follows the same compose-centric pattern as\n`vss-deploy-profile`: gather context, run preflights, work from a local copy,\ndry-run with `docker compose config`, review, deploy, then wait for health.\n\n## Standalone Deployment Flow\n\nAlways follow this sequence. Never skip the dry-run.\n\n```bash\n# 1. Copy deploy\u002Fdocker\u002Fservices\u002Frtvi\u002Frtvi-vlm\u002Frtvi-vlm-docker-compose.yml\n#    into any writable standalone working directory.\n# 2. Derive RTVI_VLM_IMAGE_TAG from that compose copy.\n# 3. Strip the standalone-only dangling depends_on block from the copy.\n# 4. Create a gitignored rtvi-vlm.env with the required RT-VLM values.\n# 5. Prepare host bind paths such as $VSS_DATA_DIR\u002Fdata_log\u002Fvst\u002Fclip_storage.\n#    Use `sudo -n` for ownership fixes; if passwordless sudo is unavailable,\n#    stop and ask the host owner to run the printed command manually.\n# 6. docker compose --env-file rtvi-vlm.env -f rtvi-vlm-docker-compose.yml config --quiet\n# 7. docker pull the exact RT-VLM image tag.\n# 8. docker compose ... up -d rtvi-vlm, wait for ready, then smoke test.\n```\n\nRun preflights before any pull or `up`; stop and fix failures here before\ndebugging RT-VLM itself:\n\n```bash\nnvidia-smi --query-gpu=index,name --format=csv,noheader\nnvidia-container-cli info\ndocker compose version\ndocker run --rm --gpus all nvidia\u002Fcuda:12.4.0-base-ubuntu22.04 nvidia-smi\n```\n\nFor standalone single-file deployments, do not run the raw\n`deploy\u002Fdocker\u002Fservices\u002Frtvi\u002Frtvi-vlm\u002Frtvi-vlm-docker-compose.yml` directly: it\ncontains `depends_on` references to sibling VLM\u002FNIM services that are only\ndefined in the full VSS\u002Fmet-blueprints compose project. The standalone reference\nshows how to copy the compose file, derive the current image tag from it, strip\nthe `depends_on` block, and validate the result before `up`.\n\nFor agent-driven validation, never let `sudo` prompt interactively. Before any\nprivileged ownership or Docker operation, use the non-interactive guard in\n[`references\u002Fdeploy-rt-vlm-service.md`](references\u002Fdeploy-rt-vlm-service.md):\nprefer plain `docker`; otherwise use `sudo -n docker`; if `sudo -n` fails, stop\nwith the exact manual command for the host owner instead of retrying with\ninteractive sudo or weakening permissions.\n\nIf `docker pull` fails with a containerd snapshotter\u002Funpack error on Docker 28+,\napply the `\u002Fetc\u002Fdocker\u002Fdaemon.json` `containerd-snapshotter=false` fix in the\nstandalone reference before retrying.\n\nMinimum standalone `rtvi-vlm.env` values:\n\n| Host env var | Required when | Purpose |\n|---|---|---|\n| `NGC_CLI_API_KEY` | Standalone deploy path | NGC registry image pull and NGC model\u002Fartifact download |\n| `RTVI_VLM_API_KEY` or `NGC_CLI_API_KEY` | Authenticated API calls | RT-VLM bearer auth after the service is running |\n| `RTVI_VLM_PORT` | Always | Host API port mapped to container `8000` |\n| `HOST_IP` | Always | Kafka bootstrap host (`${HOST_IP}:9092`) |\n| `VSS_DATA_DIR` | Always | Required clip-storage bind mount |\n| `RTVI_VLM_MODEL_TO_USE` | Always for standalone | Backend selector; use `cosmos-reason3` for the default local model or `openai-compat` for a remote\u002Fsibling endpoint |\n| `RTVI_VLM_MODEL_PATH` | Local self-hosted model | Source-backed Cosmos Reason3 Nano BF16 path: `ngc:nim\u002Fnvidia\u002Fcosmos3-nano-reasoner:bf16-final` |\n| `RTVI_VLM_ENDPOINT` | `RTVI_VLM_MODEL_TO_USE=openai-compat` | Remote\u002Fsibling OpenAI-compatible VLM endpoint |\n| `VLM_NAME` | `RTVI_VLM_MODEL_TO_USE=openai-compat` | Model\u002Fdeployment name exposed by that endpoint |\n\n## Setup\n\n```bash\nexport BASE_URL=\"http:\u002F\u002Flocalhost:${RTVI_VLM_PORT:-8018}\"  # host-side RT-VLM port\nexport API_KEY=\"${NGC_CLI_API_KEY:-${RTVI_VLM_API_KEY:-}}\" # bearer token used by host-side curl commands\n: \"${API_KEY:?Set NGC_CLI_API_KEY or RTVI_VLM_API_KEY before calling authenticated endpoints}\"\n```\n\nEvery request below uses `Authorization: Bearer $API_KEY`. Health endpoints\n(`\u002Fv1\u002Fhealth\u002F*`, `\u002Fv1\u002Fready`, `\u002Fv1\u002Flive`, `\u002Fv1\u002Fstartup`) typically work without auth.\n\n**Smoke test before use:**\n```bash\ncurl -fsS \"$BASE_URL\u002Fv1\u002Fhealth\u002Fready\"\nMODEL_ID=\"$(curl -fsS \"$BASE_URL\u002Fv1\u002Fmodels\" -H \"Authorization: Bearer $API_KEY\" | jq -r '.data[0].id \u002F\u002F .id')\"\ncurl -fsS \"$BASE_URL\u002Fopenapi.json\" | jq -r '.paths | keys[]' | sort\n```\n\n## RTSP Sample Stream Guard\n\nWhen a task or eval names `RTSP_SAMPLE_URL`, treat that exact environment\nvariable as a required input. Verify it is set and non-empty before probing or\nregistering any stream; if it is missing, stop with a clear failure message. Do\nnot derive a substitute from NvStreamer, VIOS, sample-data bundles, or any other\nfallback, because that validates a different stream than the caller requested.\n\n```bash\n: \"${RTSP_SAMPLE_URL:?Set RTSP_SAMPLE_URL to a reachable RTSP sample stream before RTSP validation}\"\ncase \"$RTSP_SAMPLE_URL\" in\n  rtsp:\u002F\u002F*) ;;\n  *) echo \"RTSP_SAMPLE_URL must be an rtsp:\u002F\u002F URL, got: $RTSP_SAMPLE_URL\" >&2; exit 1 ;;\nesac\n\nif command -v ffprobe >\u002Fdev\u002Fnull 2>&1; then\n  ffprobe -v error -rtsp_transport tcp \\\n    -select_streams v:0 -show_entries stream=codec_type \\\n    -of csv=p=0 \"$RTSP_SAMPLE_URL\" | grep -qx video\nelif command -v gst-discoverer-1.0 >\u002Fdev\u002Fnull 2>&1; then\n  gst-discoverer-1.0 \"$RTSP_SAMPLE_URL\" | grep -qi 'video'\nelse\n  echo \"Install ffprobe or gst-discoverer-1.0 before RTSP validation.\" >&2\n  exit 1\nfi\n```\n\n## Quick Start — dense captions from a local video\n\n```bash\n# 1. Upload the video, capture its file id\nFILE_ID=$(curl -fsS -X POST \"$BASE_URL\u002Fv1\u002Ffiles\" \\\n  -H \"Authorization: Bearer $API_KEY\" \\\n  -F \"file=@\u002Fpath\u002Fto\u002Fwarehouse.mp4\" \\\n  -F \"purpose=vision\" \\\n  -F \"media_type=video\" | jq -r '.id')\n\n# 2. Generate captions + alerts (SSE stream of chunked responses)\ncurl -N -X POST \"$BASE_URL\u002Fv1\u002Fgenerate_captions\" \\\n  -H \"Authorization: Bearer $API_KEY\" \\\n  -H \"Content-Type: application\u002Fjson\" \\\n  -d \"{\n    \\\"id\\\": \\\"$FILE_ID\\\",\n    \\\"prompt\\\": \\\"Write a concise dense caption for each 10-second segment of this warehouse video.\\\",\n    \\\"model\\\": \\\"$MODEL_ID\\\",\n    \\\"chunk_duration\\\": 10,\n    \\\"stream\\\": true\n  }\"\n```\n\n## API Surface\n\nUse the live OpenAPI as the source of truth before calling optional endpoints:\n\n```bash\ncurl -fsS \"$BASE_URL\u002Fopenapi.json\" | jq -r '.paths | keys[]' | sort\n```\n\nCore paths for VSS 3.2 are:\n\n- `POST \u002Fv1\u002Ffiles` for multipart media upload; pass the returned file `id` into\n  caption generation and delete the file when finished.\n- `POST \u002Fv1\u002Fgenerate_captions` for file or stream captioning. Use the exact\n  model id returned by `GET \u002Fv1\u002Fmodels`; aliases such as `cosmos-reason2` or\n  `cosmos-reason3` are backend selectors, not request model ids.\n- `POST \u002Fv1\u002Fstreams\u002Fadd`, `GET \u002Fv1\u002Fstreams\u002Fget-stream-info`, and\n  `DELETE \u002Fv1\u002Fstreams\u002Fdelete\u002F{stream_id}` for RTSP lifecycle. Parse stream ids\n  from `results[0].id`.\n- `POST \u002Fv1\u002Fchat\u002Fcompletions` for OpenAI-compatible text and multimodal calls.\n  Current 26.05 builds return HTTP 400 for text-only `\u002Fv1\u002Fcompletions`; treat\n  that as expected when validating legacy behavior.\n- `GET \u002Fv1\u002Fhealth\u002Fready`, `\u002Fv1\u002Fmodels`, `\u002Fv1\u002Fassets\u002Fstats`, and `\u002Fv1\u002Fmetrics`\n  for service probes. Do not assume `\u002Fv1\u002Flicense` exists unless OpenAPI lists it.\n\nDetailed endpoint schemas, response shapes, CV-style singular stream endpoints,\nand 26.05 compatibility notes live in\n[`references\u002Fapi-surface-26.05.md`](references\u002Fapi-surface-26.05.md).\n\n## Common Workflows\n\n- Stored file captioning: upload with `POST \u002Fv1\u002Ffiles`, call\n  `\u002Fv1\u002Fgenerate_captions` with the returned file id, use `stream=true` for SSE,\n  then delete the file to release storage.\n- RTSP live captioning: when the caller provides `RTSP_SAMPLE_URL`, use that\n  exact URL and run the **RTSP Sample Stream Guard** before registration. Do not\n  derive a replacement stream from NvStreamer or VIOS when `RTSP_SAMPLE_URL` is\n  empty; fail fast instead. Require an actual video stream\u002Fcaps entry before\n  registration; add the stream, caption it, then unregister it.\n- Alert prompts: include a deterministic `Anomaly Detected: Yes\u002FNo` line.\n  Kafka publication is server-side config, additive to HTTP responses, and\n  documented in [`references\u002Fkafka-workflows.md`](references\u002Fkafka-workflows.md).\n- Kafka validation: trust the live `vss-rtvi-vlm` environment for topic names.\n  In a full VSS alerts real-time profile, use the existing VSS Kafka container\n  `mdx-kafka` for CLI checks and final incident-consumer commands. For\n  standalone validation, use a broker that advertises `${HOST_IP}:9092`; never\n  stop or replace a pre-existing broker without user confirmation.\n\n## Error Reference\n\nCommon causes: 400 for invalid request shape or model id, 401\u002F403 for missing\nor wrong bearer token, 404 for deleted files\u002Fstreams or unsupported endpoints,\n413 for oversized uploads, 422 for schema validation, 429 for too much\nconcurrency, 500 for inference\u002Fruntime failures, and 503 while startup is still\nin progress. Inspect `docker logs vss-rtvi-vlm` for service-side failures.\n",{"data":35,"body":40},{"name":4,"description":6,"license":26,"metadata":36},{"version":37,"github-url":38,"tags":39},"3.2.1","https:\u002F\u002Fgithub.com\u002FNVIDIA-AI-Blueprints\u002Fvideo-search-and-summarization","nvidia blueprint operational deployment",{"type":41,"children":42},"root",[43,52,58,64,69,120,125,161,166,182,188,222,228,264,270,337,343,465,472,516,522,552,583,589,594,707,720,815,849,893,922,934,1208,1214,1394,1436,1444,1642,1648,1661,2154,2160,2632,2638,2643,2709,2714,2859,2874,2880,2986,2992,3005],{"type":44,"tag":45,"props":46,"children":48},"element","h2",{"id":47},"purpose",[49],{"type":50,"value":51},"text","Purpose",{"type":44,"tag":53,"props":54,"children":55},"p",{},[56],{"type":50,"value":57},"Stand up the RT-VLM dense-captioning microservice on its own and exercise every endpoint it exposes (file upload, generate_captions, stream add\u002Fdelete, chat-completions, Kafka topics).",{"type":44,"tag":45,"props":59,"children":61},{"id":60},"prerequisites",[62],{"type":50,"value":63},"Prerequisites",{"type":44,"tag":53,"props":65,"children":66},{},[67],{"type":50,"value":68},"For standalone RT-VLM deployment:",{"type":44,"tag":70,"props":71,"children":72},"ul",{},[73,79,101],{"type":44,"tag":74,"props":75,"children":76},"li",{},[77],{"type":50,"value":78},"Docker, Docker Compose, NVIDIA Container Toolkit, and a visible GPU.",{"type":44,"tag":74,"props":80,"children":81},{},[82,84,91,93,99],{"type":50,"value":83},"NGC registry credentials in ",{"type":44,"tag":85,"props":86,"children":88},"code",{"className":87},[],[89],{"type":50,"value":90},"$NGC_CLI_API_KEY",{"type":50,"value":92}," for ",{"type":44,"tag":85,"props":94,"children":96},{"className":95},[],[97],{"type":50,"value":98},"docker login nvcr.io",{"type":50,"value":100},",\nimage pulls, and local NGC model\u002Fartifact downloads.",{"type":44,"tag":74,"props":102,"children":103},{},[104,110,112,118],{"type":44,"tag":85,"props":105,"children":107},{"className":106},[],[108],{"type":50,"value":109},"curl",{"type":50,"value":111},", ",{"type":44,"tag":85,"props":113,"children":115},{"className":114},[],[116],{"type":50,"value":117},"jq",{"type":50,"value":119},", and any writable working directory for the standalone compose copy.",{"type":44,"tag":53,"props":121,"children":122},{},[123],{"type":50,"value":124},"For API calls against an existing service:",{"type":44,"tag":70,"props":126,"children":127},{},[128,141],{"type":44,"tag":74,"props":129,"children":130},{},[131,133,139],{"type":50,"value":132},"Running RT-VLM service reachable at ",{"type":44,"tag":85,"props":134,"children":136},{"className":135},[],[137],{"type":50,"value":138},"$BASE_URL",{"type":50,"value":140},".",{"type":44,"tag":74,"props":142,"children":143},{},[144,146,152,154,159],{"type":50,"value":145},"Bearer token in ",{"type":44,"tag":85,"props":147,"children":149},{"className":148},[],[150],{"type":50,"value":151},"$RTVI_VLM_API_KEY",{"type":50,"value":153}," or ",{"type":44,"tag":85,"props":155,"children":157},{"className":156},[],[158],{"type":50,"value":90},{"type":50,"value":160},", depending on how the\nservice was configured.",{"type":44,"tag":53,"props":162,"children":163},{},[164],{"type":50,"value":165},"For full VSS profile deployment:",{"type":44,"tag":70,"props":167,"children":168},{},[169],{"type":44,"tag":74,"props":170,"children":171},{},[172,174,180],{"type":50,"value":173},"Use ",{"type":44,"tag":85,"props":175,"children":177},{"className":176},[],[178],{"type":50,"value":179},"..\u002Fvss-deploy-profile\u002FSKILL.md",{"type":50,"value":181},"; this skill does not deploy full VSS profiles.",{"type":44,"tag":45,"props":183,"children":185},{"id":184},"instructions",[186],{"type":50,"value":187},"Instructions",{"type":44,"tag":53,"props":189,"children":190},{},[191,193,199,200,205,207,212,214,220],{"type":50,"value":192},"Follow the routing tables and step-by-step workflows below. Each section that ends in ",{"type":44,"tag":194,"props":195,"children":196},"em",{},[197],{"type":50,"value":198},"workflow",{"type":50,"value":111},{"type":44,"tag":194,"props":201,"children":202},{},[203],{"type":50,"value":204},"quick start",{"type":50,"value":206},", or ",{"type":44,"tag":194,"props":208,"children":209},{},[210],{"type":50,"value":211},"flow",{"type":50,"value":213}," is intended to be executed top-to-bottom. Detailed reference material lives in ",{"type":44,"tag":85,"props":215,"children":217},{"className":216},[],[218],{"type":50,"value":219},"references\u002F",{"type":50,"value":221},"; execute the documented workflows directly unless a future revision names a concrete helper.",{"type":44,"tag":45,"props":223,"children":225},{"id":224},"examples",[226],{"type":50,"value":227},"Examples",{"type":44,"tag":53,"props":229,"children":230},{},[231,233,239,241,247,249,254,256,262],{"type":50,"value":232},"Worked end-to-end examples are kept under ",{"type":44,"tag":85,"props":234,"children":236},{"className":235},[],[237],{"type":50,"value":238},"evals\u002F",{"type":50,"value":240}," (each ",{"type":44,"tag":85,"props":242,"children":244},{"className":243},[],[245],{"type":50,"value":246},"*.json",{"type":50,"value":248}," manifest contains a runnable scenario) and inline in the per-workflow ",{"type":44,"tag":85,"props":250,"children":252},{"className":251},[],[253],{"type":50,"value":109},{"type":50,"value":255}," blocks below. Run a Tier-3 evaluation with ",{"type":44,"tag":85,"props":257,"children":259},{"className":258},[],[260],{"type":50,"value":261},"nv-base validate \u003Cthis-skill-dir> --agent-eval",{"type":50,"value":263}," to replay them.",{"type":44,"tag":45,"props":265,"children":267},{"id":266},"limitations",[268],{"type":50,"value":269},"Limitations",{"type":44,"tag":70,"props":271,"children":272},{},[273,278,283,288,316],{"type":44,"tag":74,"props":274,"children":275},{},[276],{"type":50,"value":277},"Requires either a standalone RT-VLM service deployed via this skill or an\nexisting RT-VLM service reachable from the caller.",{"type":44,"tag":74,"props":279,"children":280},{},[281],{"type":50,"value":282},"NGC-hosted models and NIMs may be subject to rate-limits, GPU memory requirements, and license restrictions.",{"type":44,"tag":74,"props":284,"children":285},{},[286],{"type":50,"value":287},"Concurrency, GPU memory, and storage limits depend on the host hardware and the profile's compose file.",{"type":44,"tag":74,"props":289,"children":290},{},[291,293,299,300,306,308,314],{"type":50,"value":292},"Keep ",{"type":44,"tag":85,"props":294,"children":296},{"className":295},[],[297],{"type":50,"value":298},"NGC_CLI_API_KEY",{"type":50,"value":111},{"type":44,"tag":85,"props":301,"children":303},{"className":302},[],[304],{"type":50,"value":305},"RTVI_VLM_API_KEY",{"type":50,"value":307},", and ",{"type":44,"tag":85,"props":309,"children":311},{"className":310},[],[312],{"type":50,"value":313},"rtvi-vlm.env",{"type":50,"value":315}," files out of git and out of logs; do not echo credential values or include them in final responses.",{"type":44,"tag":74,"props":317,"children":318},{},[319,321,327,329,335],{"type":50,"value":320},"Docker group access and ",{"type":44,"tag":85,"props":322,"children":324},{"className":323},[],[325],{"type":50,"value":326},"sudo",{"type":50,"value":328}," are effectively root-level privileges. Use the non-interactive ",{"type":44,"tag":85,"props":330,"children":332},{"className":331},[],[333],{"type":50,"value":334},"sudo -n",{"type":50,"value":336}," guard in the deploy reference and stop for host-owner action when passwordless sudo is unavailable.",{"type":44,"tag":45,"props":338,"children":340},{"id":339},"troubleshooting",[341],{"type":50,"value":342},"Troubleshooting",{"type":44,"tag":70,"props":344,"children":345},{},[346,402,437],{"type":44,"tag":74,"props":347,"children":348},{},[349,355,357,362,364,369,371,377,378,384,386,392,394,400],{"type":44,"tag":350,"props":351,"children":352},"strong",{},[353],{"type":50,"value":354},"Error",{"type":50,"value":356},": REST call returns connection refused. ",{"type":44,"tag":350,"props":358,"children":359},{},[360],{"type":50,"value":361},"Cause",{"type":50,"value":363},": target microservice not running. ",{"type":44,"tag":350,"props":365,"children":366},{},[367],{"type":50,"value":368},"Solution",{"type":50,"value":370},": probe ",{"type":44,"tag":85,"props":372,"children":374},{"className":373},[],[375],{"type":50,"value":376},"\u002Fdocs",{"type":50,"value":153},{"type":44,"tag":85,"props":379,"children":381},{"className":380},[],[382],{"type":50,"value":383},"\u002Fhealth",{"type":50,"value":385},"; redeploy via ",{"type":44,"tag":85,"props":387,"children":389},{"className":388},[],[390],{"type":50,"value":391},"vss-deploy-profile",{"type":50,"value":393}," or the matching ",{"type":44,"tag":85,"props":395,"children":397},{"className":396},[],[398],{"type":50,"value":399},"vss-deploy-*",{"type":50,"value":401}," skill.",{"type":44,"tag":74,"props":403,"children":404},{},[405,409,411,415,417,422,424,428,430,435],{"type":44,"tag":350,"props":406,"children":407},{},[408],{"type":50,"value":354},{"type":50,"value":410},": HTTP 401\u002F403 from NGC pulls. ",{"type":44,"tag":350,"props":412,"children":413},{},[414],{"type":50,"value":361},{"type":50,"value":416},": missing\u002Fexpired ",{"type":44,"tag":85,"props":418,"children":420},{"className":419},[],[421],{"type":50,"value":298},{"type":50,"value":423},". ",{"type":44,"tag":350,"props":425,"children":426},{},[427],{"type":50,"value":368},{"type":50,"value":429},": ",{"type":44,"tag":85,"props":431,"children":433},{"className":432},[],[434],{"type":50,"value":98},{"type":50,"value":436}," and re-export the key before retrying.",{"type":44,"tag":74,"props":438,"children":439},{},[440,444,446,450,452,456,458,464],{"type":44,"tag":350,"props":441,"children":442},{},[443],{"type":50,"value":354},{"type":50,"value":445},": container OOM or model fails to load. ",{"type":44,"tag":350,"props":447,"children":448},{},[449],{"type":50,"value":361},{"type":50,"value":451},": insufficient GPU memory for the selected profile. ",{"type":44,"tag":350,"props":453,"children":454},{},[455],{"type":50,"value":368},{"type":50,"value":457},": switch to a smaller variant or free GPUs via ",{"type":44,"tag":85,"props":459,"children":461},{"className":460},[],[462],{"type":50,"value":463},"docker compose down",{"type":50,"value":140},{"type":44,"tag":466,"props":467,"children":469},"h1",{"id":468},"deploy-and-use-rt-vlm-dense-captioning-vss-32",[470],{"type":50,"value":471},"Deploy and Use RT-VLM Dense Captioning (VSS 3.2)",{"type":44,"tag":53,"props":473,"children":474},{},[475,477,483,484,490,491,497,499,505,507,515],{"type":50,"value":476},"RT-VLM is NVIDIA's real-time vision-language microservice: decode video (file or\nRTSP), segment it into chunks, run a VLM (",{"type":44,"tag":85,"props":478,"children":480},{"className":479},[],[481],{"type":50,"value":482},"cosmos-reason1",{"type":50,"value":111},{"type":44,"tag":85,"props":485,"children":487},{"className":486},[],[488],{"type":50,"value":489},"cosmos-reason2",{"type":50,"value":111},{"type":44,"tag":85,"props":492,"children":494},{"className":493},[],[495],{"type":50,"value":496},"cosmos-reason3",{"type":50,"value":498},", or any\nOpenAI-compatible model), stream dense captions back over SSE\u002FHTTP, and publish\ncaptions, incident alerts, and errors to Kafka. Use this skill to deploy the\nstandalone RT-VLM service when a full VSS profile is not already running, then call\nits ",{"type":44,"tag":85,"props":500,"children":502},{"className":501},[],[503],{"type":50,"value":504},"\u002Fv1\u002F...",{"type":50,"value":506}," API for caption generation, file upload, live-stream management, health\nchecks, NIM-compatible chat completions, or Prometheus metrics. API reference:\n",{"type":44,"tag":508,"props":509,"children":513},"a",{"href":510,"rel":511},"https:\u002F\u002Fdocs.nvidia.com\u002Fvss\u002Flatest\u002Freal-time-vlm-api.html",[512],"nofollow",[514],{"type":50,"value":510},{"type":50,"value":140},{"type":44,"tag":45,"props":517,"children":519},{"id":518},"deployment-routing",[520],{"type":50,"value":521},"Deployment Routing",{"type":44,"tag":53,"props":523,"children":524},{},[525,527,535,537,543,544,550],{"type":50,"value":526},"If the user asks to deploy a full VSS profile, use\n",{"type":44,"tag":508,"props":528,"children":529},{"href":179},[530],{"type":44,"tag":85,"props":531,"children":533},{"className":532},[],[534],{"type":50,"value":179},{"type":50,"value":536},". That skill\nowns profile routing, ",{"type":44,"tag":85,"props":538,"children":540},{"className":539},[],[541],{"type":50,"value":542},"generated.env",{"type":50,"value":111},{"type":44,"tag":85,"props":545,"children":547},{"className":546},[],[548],{"type":50,"value":549},"resolved.yml",{"type":50,"value":551},", multi-service sizing, and\nfull-stack deploy\u002Fteardown.",{"type":44,"tag":53,"props":553,"children":554},{},[555,557,566,568,573,575,581],{"type":50,"value":556},"If the user asks for standalone RT-VLM dense captioning, or no VSS profile is\nalready running, use the standalone RT-VLM flow in\n",{"type":44,"tag":508,"props":558,"children":560},{"href":559},"references\u002Fdeploy-rt-vlm-service.md",[561],{"type":44,"tag":85,"props":562,"children":564},{"className":563},[],[565],{"type":50,"value":559},{"type":50,"value":567},"\nbefore calling the API. This follows the same compose-centric pattern as\n",{"type":44,"tag":85,"props":569,"children":571},{"className":570},[],[572],{"type":50,"value":391},{"type":50,"value":574},": gather context, run preflights, work from a local copy,\ndry-run with ",{"type":44,"tag":85,"props":576,"children":578},{"className":577},[],[579],{"type":50,"value":580},"docker compose config",{"type":50,"value":582},", review, deploy, then wait for health.",{"type":44,"tag":45,"props":584,"children":586},{"id":585},"standalone-deployment-flow",[587],{"type":50,"value":588},"Standalone Deployment Flow",{"type":44,"tag":53,"props":590,"children":591},{},[592],{"type":50,"value":593},"Always follow this sequence. Never skip the dry-run.",{"type":44,"tag":595,"props":596,"children":601},"pre",{"className":597,"code":598,"language":599,"meta":600,"style":600},"language-bash shiki shiki-themes material-theme-lighter material-theme material-theme-palenight","# 1. Copy deploy\u002Fdocker\u002Fservices\u002Frtvi\u002Frtvi-vlm\u002Frtvi-vlm-docker-compose.yml\n#    into any writable standalone working directory.\n# 2. Derive RTVI_VLM_IMAGE_TAG from that compose copy.\n# 3. Strip the standalone-only dangling depends_on block from the copy.\n# 4. Create a gitignored rtvi-vlm.env with the required RT-VLM values.\n# 5. Prepare host bind paths such as $VSS_DATA_DIR\u002Fdata_log\u002Fvst\u002Fclip_storage.\n#    Use `sudo -n` for ownership fixes; if passwordless sudo is unavailable,\n#    stop and ask the host owner to run the printed command manually.\n# 6. docker compose --env-file rtvi-vlm.env -f rtvi-vlm-docker-compose.yml config --quiet\n# 7. docker pull the exact RT-VLM image tag.\n# 8. docker compose ... up -d rtvi-vlm, wait for ready, then smoke test.\n","bash","",[602],{"type":44,"tag":85,"props":603,"children":604},{"__ignoreMap":600},[605,617,626,635,644,653,662,671,680,689,698],{"type":44,"tag":606,"props":607,"children":610},"span",{"class":608,"line":609},"line",1,[611],{"type":44,"tag":606,"props":612,"children":614},{"style":613},"--shiki-light:#90A4AE;--shiki-light-font-style:italic;--shiki-default:#546E7A;--shiki-default-font-style:italic;--shiki-dark:#676E95;--shiki-dark-font-style:italic",[615],{"type":50,"value":616},"# 1. Copy deploy\u002Fdocker\u002Fservices\u002Frtvi\u002Frtvi-vlm\u002Frtvi-vlm-docker-compose.yml\n",{"type":44,"tag":606,"props":618,"children":620},{"class":608,"line":619},2,[621],{"type":44,"tag":606,"props":622,"children":623},{"style":613},[624],{"type":50,"value":625},"#    into any writable standalone working directory.\n",{"type":44,"tag":606,"props":627,"children":629},{"class":608,"line":628},3,[630],{"type":44,"tag":606,"props":631,"children":632},{"style":613},[633],{"type":50,"value":634},"# 2. Derive RTVI_VLM_IMAGE_TAG from that compose copy.\n",{"type":44,"tag":606,"props":636,"children":638},{"class":608,"line":637},4,[639],{"type":44,"tag":606,"props":640,"children":641},{"style":613},[642],{"type":50,"value":643},"# 3. Strip the standalone-only dangling depends_on block from the copy.\n",{"type":44,"tag":606,"props":645,"children":647},{"class":608,"line":646},5,[648],{"type":44,"tag":606,"props":649,"children":650},{"style":613},[651],{"type":50,"value":652},"# 4. Create a gitignored rtvi-vlm.env with the required RT-VLM values.\n",{"type":44,"tag":606,"props":654,"children":656},{"class":608,"line":655},6,[657],{"type":44,"tag":606,"props":658,"children":659},{"style":613},[660],{"type":50,"value":661},"# 5. Prepare host bind paths such as $VSS_DATA_DIR\u002Fdata_log\u002Fvst\u002Fclip_storage.\n",{"type":44,"tag":606,"props":663,"children":665},{"class":608,"line":664},7,[666],{"type":44,"tag":606,"props":667,"children":668},{"style":613},[669],{"type":50,"value":670},"#    Use `sudo -n` for ownership fixes; if passwordless sudo is unavailable,\n",{"type":44,"tag":606,"props":672,"children":674},{"class":608,"line":673},8,[675],{"type":44,"tag":606,"props":676,"children":677},{"style":613},[678],{"type":50,"value":679},"#    stop and ask the host owner to run the printed command manually.\n",{"type":44,"tag":606,"props":681,"children":683},{"class":608,"line":682},9,[684],{"type":44,"tag":606,"props":685,"children":686},{"style":613},[687],{"type":50,"value":688},"# 6. docker compose --env-file rtvi-vlm.env -f rtvi-vlm-docker-compose.yml config --quiet\n",{"type":44,"tag":606,"props":690,"children":692},{"class":608,"line":691},10,[693],{"type":44,"tag":606,"props":694,"children":695},{"style":613},[696],{"type":50,"value":697},"# 7. docker pull the exact RT-VLM image tag.\n",{"type":44,"tag":606,"props":699,"children":701},{"class":608,"line":700},11,[702],{"type":44,"tag":606,"props":703,"children":704},{"style":613},[705],{"type":50,"value":706},"# 8. docker compose ... up -d rtvi-vlm, wait for ready, then smoke test.\n",{"type":44,"tag":53,"props":708,"children":709},{},[710,712,718],{"type":50,"value":711},"Run preflights before any pull or ",{"type":44,"tag":85,"props":713,"children":715},{"className":714},[],[716],{"type":50,"value":717},"up",{"type":50,"value":719},"; stop and fix failures here before\ndebugging RT-VLM itself:",{"type":44,"tag":595,"props":721,"children":723},{"className":597,"code":722,"language":599,"meta":600,"style":600},"nvidia-smi --query-gpu=index,name --format=csv,noheader\nnvidia-container-cli info\ndocker compose version\ndocker run --rm --gpus all nvidia\u002Fcuda:12.4.0-base-ubuntu22.04 nvidia-smi\n",[724],{"type":44,"tag":85,"props":725,"children":726},{"__ignoreMap":600},[727,747,760,778],{"type":44,"tag":606,"props":728,"children":729},{"class":608,"line":609},[730,736,742],{"type":44,"tag":606,"props":731,"children":733},{"style":732},"--shiki-light:#E2931D;--shiki-default:#FFCB6B;--shiki-dark:#FFCB6B",[734],{"type":50,"value":735},"nvidia-smi",{"type":44,"tag":606,"props":737,"children":739},{"style":738},"--shiki-light:#91B859;--shiki-default:#C3E88D;--shiki-dark:#C3E88D",[740],{"type":50,"value":741}," --query-gpu=index,name",{"type":44,"tag":606,"props":743,"children":744},{"style":738},[745],{"type":50,"value":746}," --format=csv,noheader\n",{"type":44,"tag":606,"props":748,"children":749},{"class":608,"line":619},[750,755],{"type":44,"tag":606,"props":751,"children":752},{"style":732},[753],{"type":50,"value":754},"nvidia-container-cli",{"type":44,"tag":606,"props":756,"children":757},{"style":738},[758],{"type":50,"value":759}," info\n",{"type":44,"tag":606,"props":761,"children":762},{"class":608,"line":628},[763,768,773],{"type":44,"tag":606,"props":764,"children":765},{"style":732},[766],{"type":50,"value":767},"docker",{"type":44,"tag":606,"props":769,"children":770},{"style":738},[771],{"type":50,"value":772}," compose",{"type":44,"tag":606,"props":774,"children":775},{"style":738},[776],{"type":50,"value":777}," version\n",{"type":44,"tag":606,"props":779,"children":780},{"class":608,"line":637},[781,785,790,795,800,805,810],{"type":44,"tag":606,"props":782,"children":783},{"style":732},[784],{"type":50,"value":767},{"type":44,"tag":606,"props":786,"children":787},{"style":738},[788],{"type":50,"value":789}," run",{"type":44,"tag":606,"props":791,"children":792},{"style":738},[793],{"type":50,"value":794}," --rm",{"type":44,"tag":606,"props":796,"children":797},{"style":738},[798],{"type":50,"value":799}," --gpus",{"type":44,"tag":606,"props":801,"children":802},{"style":738},[803],{"type":50,"value":804}," all",{"type":44,"tag":606,"props":806,"children":807},{"style":738},[808],{"type":50,"value":809}," nvidia\u002Fcuda:12.4.0-base-ubuntu22.04",{"type":44,"tag":606,"props":811,"children":812},{"style":738},[813],{"type":50,"value":814}," nvidia-smi\n",{"type":44,"tag":53,"props":816,"children":817},{},[818,820,826,828,834,836,841,843,848],{"type":50,"value":819},"For standalone single-file deployments, do not run the raw\n",{"type":44,"tag":85,"props":821,"children":823},{"className":822},[],[824],{"type":50,"value":825},"deploy\u002Fdocker\u002Fservices\u002Frtvi\u002Frtvi-vlm\u002Frtvi-vlm-docker-compose.yml",{"type":50,"value":827}," directly: it\ncontains ",{"type":44,"tag":85,"props":829,"children":831},{"className":830},[],[832],{"type":50,"value":833},"depends_on",{"type":50,"value":835}," references to sibling VLM\u002FNIM services that are only\ndefined in the full VSS\u002Fmet-blueprints compose project. The standalone reference\nshows how to copy the compose file, derive the current image tag from it, strip\nthe ",{"type":44,"tag":85,"props":837,"children":839},{"className":838},[],[840],{"type":50,"value":833},{"type":50,"value":842}," block, and validate the result before ",{"type":44,"tag":85,"props":844,"children":846},{"className":845},[],[847],{"type":50,"value":717},{"type":50,"value":140},{"type":44,"tag":53,"props":850,"children":851},{},[852,854,859,861,869,871,876,878,884,886,891],{"type":50,"value":853},"For agent-driven validation, never let ",{"type":44,"tag":85,"props":855,"children":857},{"className":856},[],[858],{"type":50,"value":326},{"type":50,"value":860}," prompt interactively. Before any\nprivileged ownership or Docker operation, use the non-interactive guard in\n",{"type":44,"tag":508,"props":862,"children":863},{"href":559},[864],{"type":44,"tag":85,"props":865,"children":867},{"className":866},[],[868],{"type":50,"value":559},{"type":50,"value":870},":\nprefer plain ",{"type":44,"tag":85,"props":872,"children":874},{"className":873},[],[875],{"type":50,"value":767},{"type":50,"value":877},"; otherwise use ",{"type":44,"tag":85,"props":879,"children":881},{"className":880},[],[882],{"type":50,"value":883},"sudo -n docker",{"type":50,"value":885},"; if ",{"type":44,"tag":85,"props":887,"children":889},{"className":888},[],[890],{"type":50,"value":334},{"type":50,"value":892}," fails, stop\nwith the exact manual command for the host owner instead of retrying with\ninteractive sudo or weakening permissions.",{"type":44,"tag":53,"props":894,"children":895},{},[896,898,904,906,912,914,920],{"type":50,"value":897},"If ",{"type":44,"tag":85,"props":899,"children":901},{"className":900},[],[902],{"type":50,"value":903},"docker pull",{"type":50,"value":905}," fails with a containerd snapshotter\u002Funpack error on Docker 28+,\napply the ",{"type":44,"tag":85,"props":907,"children":909},{"className":908},[],[910],{"type":50,"value":911},"\u002Fetc\u002Fdocker\u002Fdaemon.json",{"type":50,"value":913}," ",{"type":44,"tag":85,"props":915,"children":917},{"className":916},[],[918],{"type":50,"value":919},"containerd-snapshotter=false",{"type":50,"value":921}," fix in the\nstandalone reference before retrying.",{"type":44,"tag":53,"props":923,"children":924},{},[925,927,932],{"type":50,"value":926},"Minimum standalone ",{"type":44,"tag":85,"props":928,"children":930},{"className":929},[],[931],{"type":50,"value":313},{"type":50,"value":933}," values:",{"type":44,"tag":935,"props":936,"children":937},"table",{},[938,961],{"type":44,"tag":939,"props":940,"children":941},"thead",{},[942],{"type":44,"tag":943,"props":944,"children":945},"tr",{},[946,952,957],{"type":44,"tag":947,"props":948,"children":949},"th",{},[950],{"type":50,"value":951},"Host env var",{"type":44,"tag":947,"props":953,"children":954},{},[955],{"type":50,"value":956},"Required when",{"type":44,"tag":947,"props":958,"children":959},{},[960],{"type":50,"value":51},{"type":44,"tag":962,"props":963,"children":964},"tbody",{},[965,987,1014,1042,1071,1092,1129,1157,1183],{"type":44,"tag":943,"props":966,"children":967},{},[968,977,982],{"type":44,"tag":969,"props":970,"children":971},"td",{},[972],{"type":44,"tag":85,"props":973,"children":975},{"className":974},[],[976],{"type":50,"value":298},{"type":44,"tag":969,"props":978,"children":979},{},[980],{"type":50,"value":981},"Standalone deploy path",{"type":44,"tag":969,"props":983,"children":984},{},[985],{"type":50,"value":986},"NGC registry image pull and NGC model\u002Fartifact download",{"type":44,"tag":943,"props":988,"children":989},{},[990,1004,1009],{"type":44,"tag":969,"props":991,"children":992},{},[993,998,999],{"type":44,"tag":85,"props":994,"children":996},{"className":995},[],[997],{"type":50,"value":305},{"type":50,"value":153},{"type":44,"tag":85,"props":1000,"children":1002},{"className":1001},[],[1003],{"type":50,"value":298},{"type":44,"tag":969,"props":1005,"children":1006},{},[1007],{"type":50,"value":1008},"Authenticated API calls",{"type":44,"tag":969,"props":1010,"children":1011},{},[1012],{"type":50,"value":1013},"RT-VLM bearer auth after the service is running",{"type":44,"tag":943,"props":1015,"children":1016},{},[1017,1026,1031],{"type":44,"tag":969,"props":1018,"children":1019},{},[1020],{"type":44,"tag":85,"props":1021,"children":1023},{"className":1022},[],[1024],{"type":50,"value":1025},"RTVI_VLM_PORT",{"type":44,"tag":969,"props":1027,"children":1028},{},[1029],{"type":50,"value":1030},"Always",{"type":44,"tag":969,"props":1032,"children":1033},{},[1034,1036],{"type":50,"value":1035},"Host API port mapped to container ",{"type":44,"tag":85,"props":1037,"children":1039},{"className":1038},[],[1040],{"type":50,"value":1041},"8000",{"type":44,"tag":943,"props":1043,"children":1044},{},[1045,1054,1058],{"type":44,"tag":969,"props":1046,"children":1047},{},[1048],{"type":44,"tag":85,"props":1049,"children":1051},{"className":1050},[],[1052],{"type":50,"value":1053},"HOST_IP",{"type":44,"tag":969,"props":1055,"children":1056},{},[1057],{"type":50,"value":1030},{"type":44,"tag":969,"props":1059,"children":1060},{},[1061,1063,1069],{"type":50,"value":1062},"Kafka bootstrap host (",{"type":44,"tag":85,"props":1064,"children":1066},{"className":1065},[],[1067],{"type":50,"value":1068},"${HOST_IP}:9092",{"type":50,"value":1070},")",{"type":44,"tag":943,"props":1072,"children":1073},{},[1074,1083,1087],{"type":44,"tag":969,"props":1075,"children":1076},{},[1077],{"type":44,"tag":85,"props":1078,"children":1080},{"className":1079},[],[1081],{"type":50,"value":1082},"VSS_DATA_DIR",{"type":44,"tag":969,"props":1084,"children":1085},{},[1086],{"type":50,"value":1030},{"type":44,"tag":969,"props":1088,"children":1089},{},[1090],{"type":50,"value":1091},"Required clip-storage bind mount",{"type":44,"tag":943,"props":1093,"children":1094},{},[1095,1104,1109],{"type":44,"tag":969,"props":1096,"children":1097},{},[1098],{"type":44,"tag":85,"props":1099,"children":1101},{"className":1100},[],[1102],{"type":50,"value":1103},"RTVI_VLM_MODEL_TO_USE",{"type":44,"tag":969,"props":1105,"children":1106},{},[1107],{"type":50,"value":1108},"Always for standalone",{"type":44,"tag":969,"props":1110,"children":1111},{},[1112,1114,1119,1121,1127],{"type":50,"value":1113},"Backend selector; use ",{"type":44,"tag":85,"props":1115,"children":1117},{"className":1116},[],[1118],{"type":50,"value":496},{"type":50,"value":1120}," for the default local model or ",{"type":44,"tag":85,"props":1122,"children":1124},{"className":1123},[],[1125],{"type":50,"value":1126},"openai-compat",{"type":50,"value":1128}," for a remote\u002Fsibling endpoint",{"type":44,"tag":943,"props":1130,"children":1131},{},[1132,1141,1146],{"type":44,"tag":969,"props":1133,"children":1134},{},[1135],{"type":44,"tag":85,"props":1136,"children":1138},{"className":1137},[],[1139],{"type":50,"value":1140},"RTVI_VLM_MODEL_PATH",{"type":44,"tag":969,"props":1142,"children":1143},{},[1144],{"type":50,"value":1145},"Local self-hosted model",{"type":44,"tag":969,"props":1147,"children":1148},{},[1149,1151],{"type":50,"value":1150},"Source-backed Cosmos Reason3 Nano BF16 path: ",{"type":44,"tag":85,"props":1152,"children":1154},{"className":1153},[],[1155],{"type":50,"value":1156},"ngc:nim\u002Fnvidia\u002Fcosmos3-nano-reasoner:bf16-final",{"type":44,"tag":943,"props":1158,"children":1159},{},[1160,1169,1178],{"type":44,"tag":969,"props":1161,"children":1162},{},[1163],{"type":44,"tag":85,"props":1164,"children":1166},{"className":1165},[],[1167],{"type":50,"value":1168},"RTVI_VLM_ENDPOINT",{"type":44,"tag":969,"props":1170,"children":1171},{},[1172],{"type":44,"tag":85,"props":1173,"children":1175},{"className":1174},[],[1176],{"type":50,"value":1177},"RTVI_VLM_MODEL_TO_USE=openai-compat",{"type":44,"tag":969,"props":1179,"children":1180},{},[1181],{"type":50,"value":1182},"Remote\u002Fsibling OpenAI-compatible VLM endpoint",{"type":44,"tag":943,"props":1184,"children":1185},{},[1186,1195,1203],{"type":44,"tag":969,"props":1187,"children":1188},{},[1189],{"type":44,"tag":85,"props":1190,"children":1192},{"className":1191},[],[1193],{"type":50,"value":1194},"VLM_NAME",{"type":44,"tag":969,"props":1196,"children":1197},{},[1198],{"type":44,"tag":85,"props":1199,"children":1201},{"className":1200},[],[1202],{"type":50,"value":1177},{"type":44,"tag":969,"props":1204,"children":1205},{},[1206],{"type":50,"value":1207},"Model\u002Fdeployment name exposed by that endpoint",{"type":44,"tag":45,"props":1209,"children":1211},{"id":1210},"setup",[1212],{"type":50,"value":1213},"Setup",{"type":44,"tag":595,"props":1215,"children":1217},{"className":597,"code":1216,"language":599,"meta":600,"style":600},"export BASE_URL=\"http:\u002F\u002Flocalhost:${RTVI_VLM_PORT:-8018}\"  # host-side RT-VLM port\nexport API_KEY=\"${NGC_CLI_API_KEY:-${RTVI_VLM_API_KEY:-}}\" # bearer token used by host-side curl commands\n: \"${API_KEY:?Set NGC_CLI_API_KEY or RTVI_VLM_API_KEY before calling authenticated endpoints}\"\n",[1218],{"type":44,"tag":85,"props":1219,"children":1220},{"__ignoreMap":600},[1221,1281,1325],{"type":44,"tag":606,"props":1222,"children":1223},{"class":608,"line":609},[1224,1230,1236,1242,1247,1252,1257,1261,1266,1271,1276],{"type":44,"tag":606,"props":1225,"children":1227},{"style":1226},"--shiki-light:#9C3EDA;--shiki-default:#C792EA;--shiki-dark:#C792EA",[1228],{"type":50,"value":1229},"export",{"type":44,"tag":606,"props":1231,"children":1233},{"style":1232},"--shiki-light:#90A4AE;--shiki-default:#EEFFFF;--shiki-dark:#BABED8",[1234],{"type":50,"value":1235}," BASE_URL",{"type":44,"tag":606,"props":1237,"children":1239},{"style":1238},"--shiki-light:#39ADB5;--shiki-default:#89DDFF;--shiki-dark:#89DDFF",[1240],{"type":50,"value":1241},"=",{"type":44,"tag":606,"props":1243,"children":1244},{"style":1238},[1245],{"type":50,"value":1246},"\"",{"type":44,"tag":606,"props":1248,"children":1249},{"style":738},[1250],{"type":50,"value":1251},"http:\u002F\u002Flocalhost:",{"type":44,"tag":606,"props":1253,"children":1254},{"style":1238},[1255],{"type":50,"value":1256},"${",{"type":44,"tag":606,"props":1258,"children":1259},{"style":1232},[1260],{"type":50,"value":1025},{"type":44,"tag":606,"props":1262,"children":1263},{"style":1238},[1264],{"type":50,"value":1265},":-",{"type":44,"tag":606,"props":1267,"children":1268},{"style":1232},[1269],{"type":50,"value":1270},"8018",{"type":44,"tag":606,"props":1272,"children":1273},{"style":1238},[1274],{"type":50,"value":1275},"}\"",{"type":44,"tag":606,"props":1277,"children":1278},{"style":613},[1279],{"type":50,"value":1280},"  # host-side RT-VLM port\n",{"type":44,"tag":606,"props":1282,"children":1283},{"class":608,"line":619},[1284,1288,1293,1297,1302,1306,1311,1315,1320],{"type":44,"tag":606,"props":1285,"children":1286},{"style":1226},[1287],{"type":50,"value":1229},{"type":44,"tag":606,"props":1289,"children":1290},{"style":1232},[1291],{"type":50,"value":1292}," API_KEY",{"type":44,"tag":606,"props":1294,"children":1295},{"style":1238},[1296],{"type":50,"value":1241},{"type":44,"tag":606,"props":1298,"children":1299},{"style":1238},[1300],{"type":50,"value":1301},"\"${",{"type":44,"tag":606,"props":1303,"children":1304},{"style":1232},[1305],{"type":50,"value":298},{"type":44,"tag":606,"props":1307,"children":1308},{"style":1238},[1309],{"type":50,"value":1310},":-${",{"type":44,"tag":606,"props":1312,"children":1313},{"style":1232},[1314],{"type":50,"value":305},{"type":44,"tag":606,"props":1316,"children":1317},{"style":1238},[1318],{"type":50,"value":1319},":-}}\"",{"type":44,"tag":606,"props":1321,"children":1322},{"style":613},[1323],{"type":50,"value":1324}," # bearer token used by host-side curl commands\n",{"type":44,"tag":606,"props":1326,"children":1327},{"class":608,"line":628},[1328,1334,1339,1344,1349,1354,1359,1364,1369,1374,1379,1384,1389],{"type":44,"tag":606,"props":1329,"children":1331},{"style":1330},"--shiki-light:#6182B8;--shiki-default:#82AAFF;--shiki-dark:#82AAFF",[1332],{"type":50,"value":1333},":",{"type":44,"tag":606,"props":1335,"children":1336},{"style":1238},[1337],{"type":50,"value":1338}," \"${",{"type":44,"tag":606,"props":1340,"children":1341},{"style":1232},[1342],{"type":50,"value":1343},"API_KEY",{"type":44,"tag":606,"props":1345,"children":1346},{"style":1238},[1347],{"type":50,"value":1348},":?",{"type":44,"tag":606,"props":1350,"children":1351},{"style":1232},[1352],{"type":50,"value":1353},"Set",{"type":44,"tag":606,"props":1355,"children":1356},{"style":1232},[1357],{"type":50,"value":1358}," NGC_CLI_API_KEY",{"type":44,"tag":606,"props":1360,"children":1361},{"style":1232},[1362],{"type":50,"value":1363}," or",{"type":44,"tag":606,"props":1365,"children":1366},{"style":1232},[1367],{"type":50,"value":1368}," RTVI_VLM_API_KEY",{"type":44,"tag":606,"props":1370,"children":1371},{"style":1232},[1372],{"type":50,"value":1373}," before",{"type":44,"tag":606,"props":1375,"children":1376},{"style":1232},[1377],{"type":50,"value":1378}," calling",{"type":44,"tag":606,"props":1380,"children":1381},{"style":1232},[1382],{"type":50,"value":1383}," authenticated",{"type":44,"tag":606,"props":1385,"children":1386},{"style":1232},[1387],{"type":50,"value":1388}," endpoints",{"type":44,"tag":606,"props":1390,"children":1391},{"style":1238},[1392],{"type":50,"value":1393},"}\"\n",{"type":44,"tag":53,"props":1395,"children":1396},{},[1397,1399,1405,1407,1413,1414,1420,1421,1427,1428,1434],{"type":50,"value":1398},"Every request below uses ",{"type":44,"tag":85,"props":1400,"children":1402},{"className":1401},[],[1403],{"type":50,"value":1404},"Authorization: Bearer $API_KEY",{"type":50,"value":1406},". Health endpoints\n(",{"type":44,"tag":85,"props":1408,"children":1410},{"className":1409},[],[1411],{"type":50,"value":1412},"\u002Fv1\u002Fhealth\u002F*",{"type":50,"value":111},{"type":44,"tag":85,"props":1415,"children":1417},{"className":1416},[],[1418],{"type":50,"value":1419},"\u002Fv1\u002Fready",{"type":50,"value":111},{"type":44,"tag":85,"props":1422,"children":1424},{"className":1423},[],[1425],{"type":50,"value":1426},"\u002Fv1\u002Flive",{"type":50,"value":111},{"type":44,"tag":85,"props":1429,"children":1431},{"className":1430},[],[1432],{"type":50,"value":1433},"\u002Fv1\u002Fstartup",{"type":50,"value":1435},") typically work without auth.",{"type":44,"tag":53,"props":1437,"children":1438},{},[1439],{"type":44,"tag":350,"props":1440,"children":1441},{},[1442],{"type":50,"value":1443},"Smoke test before use:",{"type":44,"tag":595,"props":1445,"children":1447},{"className":597,"code":1446,"language":599,"meta":600,"style":600},"curl -fsS \"$BASE_URL\u002Fv1\u002Fhealth\u002Fready\"\nMODEL_ID=\"$(curl -fsS \"$BASE_URL\u002Fv1\u002Fmodels\" -H \"Authorization: Bearer $API_KEY\" | jq -r '.data[0].id \u002F\u002F .id')\"\ncurl -fsS \"$BASE_URL\u002Fopenapi.json\" | jq -r '.paths | keys[]' | sort\n",[1448],{"type":44,"tag":85,"props":1449,"children":1450},{"__ignoreMap":600},[1451,1482,1578],{"type":44,"tag":606,"props":1452,"children":1453},{"class":608,"line":609},[1454,1458,1463,1468,1472,1477],{"type":44,"tag":606,"props":1455,"children":1456},{"style":732},[1457],{"type":50,"value":109},{"type":44,"tag":606,"props":1459,"children":1460},{"style":738},[1461],{"type":50,"value":1462}," -fsS",{"type":44,"tag":606,"props":1464,"children":1465},{"style":1238},[1466],{"type":50,"value":1467}," \"",{"type":44,"tag":606,"props":1469,"children":1470},{"style":1232},[1471],{"type":50,"value":138},{"type":44,"tag":606,"props":1473,"children":1474},{"style":738},[1475],{"type":50,"value":1476},"\u002Fv1\u002Fhealth\u002Fready",{"type":44,"tag":606,"props":1478,"children":1479},{"style":1238},[1480],{"type":50,"value":1481},"\"\n",{"type":44,"tag":606,"props":1483,"children":1484},{"class":608,"line":619},[1485,1490,1494,1499,1503,1508,1512,1516,1521,1525,1530,1534,1539,1544,1548,1553,1558,1563,1568,1573],{"type":44,"tag":606,"props":1486,"children":1487},{"style":1232},[1488],{"type":50,"value":1489},"MODEL_ID",{"type":44,"tag":606,"props":1491,"children":1492},{"style":1238},[1493],{"type":50,"value":1241},{"type":44,"tag":606,"props":1495,"children":1496},{"style":1238},[1497],{"type":50,"value":1498},"\"$(",{"type":44,"tag":606,"props":1500,"children":1501},{"style":732},[1502],{"type":50,"value":109},{"type":44,"tag":606,"props":1504,"children":1505},{"style":738},[1506],{"type":50,"value":1507}," -fsS ",{"type":44,"tag":606,"props":1509,"children":1510},{"style":1238},[1511],{"type":50,"value":1246},{"type":44,"tag":606,"props":1513,"children":1514},{"style":1232},[1515],{"type":50,"value":138},{"type":44,"tag":606,"props":1517,"children":1518},{"style":738},[1519],{"type":50,"value":1520},"\u002Fv1\u002Fmodels",{"type":44,"tag":606,"props":1522,"children":1523},{"style":1238},[1524],{"type":50,"value":1246},{"type":44,"tag":606,"props":1526,"children":1527},{"style":738},[1528],{"type":50,"value":1529}," -H ",{"type":44,"tag":606,"props":1531,"children":1532},{"style":1238},[1533],{"type":50,"value":1246},{"type":44,"tag":606,"props":1535,"children":1536},{"style":738},[1537],{"type":50,"value":1538},"Authorization: Bearer ",{"type":44,"tag":606,"props":1540,"children":1541},{"style":1232},[1542],{"type":50,"value":1543},"$API_KEY",{"type":44,"tag":606,"props":1545,"children":1546},{"style":1238},[1547],{"type":50,"value":1246},{"type":44,"tag":606,"props":1549,"children":1550},{"style":1238},[1551],{"type":50,"value":1552}," |",{"type":44,"tag":606,"props":1554,"children":1555},{"style":732},[1556],{"type":50,"value":1557}," jq",{"type":44,"tag":606,"props":1559,"children":1560},{"style":738},[1561],{"type":50,"value":1562}," -r ",{"type":44,"tag":606,"props":1564,"children":1565},{"style":1238},[1566],{"type":50,"value":1567},"'",{"type":44,"tag":606,"props":1569,"children":1570},{"style":738},[1571],{"type":50,"value":1572},".data[0].id \u002F\u002F .id",{"type":44,"tag":606,"props":1574,"children":1575},{"style":1238},[1576],{"type":50,"value":1577},"')\"\n",{"type":44,"tag":606,"props":1579,"children":1580},{"class":608,"line":628},[1581,1585,1589,1593,1597,1602,1606,1610,1614,1619,1624,1629,1633,1637],{"type":44,"tag":606,"props":1582,"children":1583},{"style":732},[1584],{"type":50,"value":109},{"type":44,"tag":606,"props":1586,"children":1587},{"style":738},[1588],{"type":50,"value":1462},{"type":44,"tag":606,"props":1590,"children":1591},{"style":1238},[1592],{"type":50,"value":1467},{"type":44,"tag":606,"props":1594,"children":1595},{"style":1232},[1596],{"type":50,"value":138},{"type":44,"tag":606,"props":1598,"children":1599},{"style":738},[1600],{"type":50,"value":1601},"\u002Fopenapi.json",{"type":44,"tag":606,"props":1603,"children":1604},{"style":1238},[1605],{"type":50,"value":1246},{"type":44,"tag":606,"props":1607,"children":1608},{"style":1238},[1609],{"type":50,"value":1552},{"type":44,"tag":606,"props":1611,"children":1612},{"style":732},[1613],{"type":50,"value":1557},{"type":44,"tag":606,"props":1615,"children":1616},{"style":738},[1617],{"type":50,"value":1618}," -r",{"type":44,"tag":606,"props":1620,"children":1621},{"style":1238},[1622],{"type":50,"value":1623}," '",{"type":44,"tag":606,"props":1625,"children":1626},{"style":738},[1627],{"type":50,"value":1628},".paths | keys[]",{"type":44,"tag":606,"props":1630,"children":1631},{"style":1238},[1632],{"type":50,"value":1567},{"type":44,"tag":606,"props":1634,"children":1635},{"style":1238},[1636],{"type":50,"value":1552},{"type":44,"tag":606,"props":1638,"children":1639},{"style":732},[1640],{"type":50,"value":1641}," sort\n",{"type":44,"tag":45,"props":1643,"children":1645},{"id":1644},"rtsp-sample-stream-guard",[1646],{"type":50,"value":1647},"RTSP Sample Stream Guard",{"type":44,"tag":53,"props":1649,"children":1650},{},[1651,1653,1659],{"type":50,"value":1652},"When a task or eval names ",{"type":44,"tag":85,"props":1654,"children":1656},{"className":1655},[],[1657],{"type":50,"value":1658},"RTSP_SAMPLE_URL",{"type":50,"value":1660},", treat that exact environment\nvariable as a required input. Verify it is set and non-empty before probing or\nregistering any stream; if it is missing, stop with a clear failure message. Do\nnot derive a substitute from NvStreamer, VIOS, sample-data bundles, or any other\nfallback, because that validates a different stream than the caller requested.",{"type":44,"tag":595,"props":1662,"children":1664},{"className":597,"code":1663,"language":599,"meta":600,"style":600},": \"${RTSP_SAMPLE_URL:?Set RTSP_SAMPLE_URL to a reachable RTSP sample stream before RTSP validation}\"\ncase \"$RTSP_SAMPLE_URL\" in\n  rtsp:\u002F\u002F*) ;;\n  *) echo \"RTSP_SAMPLE_URL must be an rtsp:\u002F\u002F URL, got: $RTSP_SAMPLE_URL\" >&2; exit 1 ;;\nesac\n\nif command -v ffprobe >\u002Fdev\u002Fnull 2>&1; then\n  ffprobe -v error -rtsp_transport tcp \\\n    -select_streams v:0 -show_entries stream=codec_type \\\n    -of csv=p=0 \"$RTSP_SAMPLE_URL\" | grep -qx video\nelif command -v gst-discoverer-1.0 >\u002Fdev\u002Fnull 2>&1; then\n  gst-discoverer-1.0 \"$RTSP_SAMPLE_URL\" | grep -qi 'video'\nelse\n  echo \"Install ffprobe or gst-discoverer-1.0 before RTSP validation.\" >&2\n  exit 1\nfi\n",[1665],{"type":44,"tag":85,"props":1666,"children":1667},{"__ignoreMap":600},[1668,1743,1770,1792,1842,1850,1859,1902,1934,1961,2010,2047,2095,2104,2131,2145],{"type":44,"tag":606,"props":1669,"children":1670},{"class":608,"line":609},[1671,1675,1679,1683,1687,1691,1696,1701,1706,1711,1716,1721,1726,1730,1734,1739],{"type":44,"tag":606,"props":1672,"children":1673},{"style":1330},[1674],{"type":50,"value":1333},{"type":44,"tag":606,"props":1676,"children":1677},{"style":1238},[1678],{"type":50,"value":1338},{"type":44,"tag":606,"props":1680,"children":1681},{"style":1232},[1682],{"type":50,"value":1658},{"type":44,"tag":606,"props":1684,"children":1685},{"style":1238},[1686],{"type":50,"value":1348},{"type":44,"tag":606,"props":1688,"children":1689},{"style":1232},[1690],{"type":50,"value":1353},{"type":44,"tag":606,"props":1692,"children":1693},{"style":1232},[1694],{"type":50,"value":1695}," RTSP_SAMPLE_URL",{"type":44,"tag":606,"props":1697,"children":1698},{"style":1232},[1699],{"type":50,"value":1700}," to",{"type":44,"tag":606,"props":1702,"children":1703},{"style":1232},[1704],{"type":50,"value":1705}," a",{"type":44,"tag":606,"props":1707,"children":1708},{"style":1232},[1709],{"type":50,"value":1710}," reachable",{"type":44,"tag":606,"props":1712,"children":1713},{"style":1232},[1714],{"type":50,"value":1715}," RTSP",{"type":44,"tag":606,"props":1717,"children":1718},{"style":1232},[1719],{"type":50,"value":1720}," sample",{"type":44,"tag":606,"props":1722,"children":1723},{"style":1232},[1724],{"type":50,"value":1725}," stream",{"type":44,"tag":606,"props":1727,"children":1728},{"style":1232},[1729],{"type":50,"value":1373},{"type":44,"tag":606,"props":1731,"children":1732},{"style":1232},[1733],{"type":50,"value":1715},{"type":44,"tag":606,"props":1735,"children":1736},{"style":1232},[1737],{"type":50,"value":1738}," validation",{"type":44,"tag":606,"props":1740,"children":1741},{"style":1238},[1742],{"type":50,"value":1393},{"type":44,"tag":606,"props":1744,"children":1745},{"class":608,"line":619},[1746,1752,1756,1761,1765],{"type":44,"tag":606,"props":1747,"children":1749},{"style":1748},"--shiki-light:#39ADB5;--shiki-light-font-style:italic;--shiki-default:#89DDFF;--shiki-default-font-style:italic;--shiki-dark:#89DDFF;--shiki-dark-font-style:italic",[1750],{"type":50,"value":1751},"case",{"type":44,"tag":606,"props":1753,"children":1754},{"style":1238},[1755],{"type":50,"value":1467},{"type":44,"tag":606,"props":1757,"children":1758},{"style":1232},[1759],{"type":50,"value":1760},"$RTSP_SAMPLE_URL",{"type":44,"tag":606,"props":1762,"children":1763},{"style":1238},[1764],{"type":50,"value":1246},{"type":44,"tag":606,"props":1766,"children":1767},{"style":1748},[1768],{"type":50,"value":1769}," in\n",{"type":44,"tag":606,"props":1771,"children":1772},{"class":608,"line":628},[1773,1778,1783,1787],{"type":44,"tag":606,"props":1774,"children":1775},{"style":738},[1776],{"type":50,"value":1777},"  rtsp:\u002F\u002F",{"type":44,"tag":606,"props":1779,"children":1780},{"style":1238},[1781],{"type":50,"value":1782},"*",{"type":44,"tag":606,"props":1784,"children":1785},{"style":1238},[1786],{"type":50,"value":1070},{"type":44,"tag":606,"props":1788,"children":1789},{"style":1238},[1790],{"type":50,"value":1791}," ;;\n",{"type":44,"tag":606,"props":1793,"children":1794},{"class":608,"line":637},[1795,1800,1805,1809,1814,1818,1822,1827,1832,1838],{"type":44,"tag":606,"props":1796,"children":1797},{"style":1238},[1798],{"type":50,"value":1799},"  *)",{"type":44,"tag":606,"props":1801,"children":1802},{"style":1330},[1803],{"type":50,"value":1804}," echo",{"type":44,"tag":606,"props":1806,"children":1807},{"style":1238},[1808],{"type":50,"value":1467},{"type":44,"tag":606,"props":1810,"children":1811},{"style":738},[1812],{"type":50,"value":1813},"RTSP_SAMPLE_URL must be an rtsp:\u002F\u002F URL, got: ",{"type":44,"tag":606,"props":1815,"children":1816},{"style":1232},[1817],{"type":50,"value":1760},{"type":44,"tag":606,"props":1819,"children":1820},{"style":1238},[1821],{"type":50,"value":1246},{"type":44,"tag":606,"props":1823,"children":1824},{"style":1238},[1825],{"type":50,"value":1826}," >&2;",{"type":44,"tag":606,"props":1828,"children":1829},{"style":1330},[1830],{"type":50,"value":1831}," exit",{"type":44,"tag":606,"props":1833,"children":1835},{"style":1834},"--shiki-light:#F76D47;--shiki-default:#F78C6C;--shiki-dark:#F78C6C",[1836],{"type":50,"value":1837}," 1",{"type":44,"tag":606,"props":1839,"children":1840},{"style":1238},[1841],{"type":50,"value":1791},{"type":44,"tag":606,"props":1843,"children":1844},{"class":608,"line":646},[1845],{"type":44,"tag":606,"props":1846,"children":1847},{"style":1748},[1848],{"type":50,"value":1849},"esac\n",{"type":44,"tag":606,"props":1851,"children":1852},{"class":608,"line":655},[1853],{"type":44,"tag":606,"props":1854,"children":1856},{"emptyLinePlaceholder":1855},true,[1857],{"type":50,"value":1858},"\n",{"type":44,"tag":606,"props":1860,"children":1861},{"class":608,"line":664},[1862,1867,1872,1877,1882,1887,1892,1897],{"type":44,"tag":606,"props":1863,"children":1864},{"style":1748},[1865],{"type":50,"value":1866},"if",{"type":44,"tag":606,"props":1868,"children":1869},{"style":1330},[1870],{"type":50,"value":1871}," command",{"type":44,"tag":606,"props":1873,"children":1874},{"style":738},[1875],{"type":50,"value":1876}," -v",{"type":44,"tag":606,"props":1878,"children":1879},{"style":738},[1880],{"type":50,"value":1881}," ffprobe",{"type":44,"tag":606,"props":1883,"children":1884},{"style":1238},[1885],{"type":50,"value":1886}," >",{"type":44,"tag":606,"props":1888,"children":1889},{"style":738},[1890],{"type":50,"value":1891},"\u002Fdev\u002Fnull",{"type":44,"tag":606,"props":1893,"children":1894},{"style":1238},[1895],{"type":50,"value":1896}," 2>&1;",{"type":44,"tag":606,"props":1898,"children":1899},{"style":1748},[1900],{"type":50,"value":1901}," then\n",{"type":44,"tag":606,"props":1903,"children":1904},{"class":608,"line":673},[1905,1910,1914,1919,1924,1929],{"type":44,"tag":606,"props":1906,"children":1907},{"style":732},[1908],{"type":50,"value":1909},"  ffprobe",{"type":44,"tag":606,"props":1911,"children":1912},{"style":738},[1913],{"type":50,"value":1876},{"type":44,"tag":606,"props":1915,"children":1916},{"style":738},[1917],{"type":50,"value":1918}," error",{"type":44,"tag":606,"props":1920,"children":1921},{"style":738},[1922],{"type":50,"value":1923}," -rtsp_transport",{"type":44,"tag":606,"props":1925,"children":1926},{"style":738},[1927],{"type":50,"value":1928}," tcp",{"type":44,"tag":606,"props":1930,"children":1931},{"style":1232},[1932],{"type":50,"value":1933}," \\\n",{"type":44,"tag":606,"props":1935,"children":1936},{"class":608,"line":682},[1937,1942,1947,1952,1957],{"type":44,"tag":606,"props":1938,"children":1939},{"style":738},[1940],{"type":50,"value":1941},"    -select_streams",{"type":44,"tag":606,"props":1943,"children":1944},{"style":738},[1945],{"type":50,"value":1946}," v:0",{"type":44,"tag":606,"props":1948,"children":1949},{"style":738},[1950],{"type":50,"value":1951}," -show_entries",{"type":44,"tag":606,"props":1953,"children":1954},{"style":738},[1955],{"type":50,"value":1956}," stream=codec_type",{"type":44,"tag":606,"props":1958,"children":1959},{"style":1232},[1960],{"type":50,"value":1933},{"type":44,"tag":606,"props":1962,"children":1963},{"class":608,"line":691},[1964,1969,1974,1979,1983,1987,1991,1995,2000,2005],{"type":44,"tag":606,"props":1965,"children":1966},{"style":738},[1967],{"type":50,"value":1968},"    -of",{"type":44,"tag":606,"props":1970,"children":1971},{"style":738},[1972],{"type":50,"value":1973}," csv=p=",{"type":44,"tag":606,"props":1975,"children":1976},{"style":1834},[1977],{"type":50,"value":1978},"0",{"type":44,"tag":606,"props":1980,"children":1981},{"style":1238},[1982],{"type":50,"value":1467},{"type":44,"tag":606,"props":1984,"children":1985},{"style":1232},[1986],{"type":50,"value":1760},{"type":44,"tag":606,"props":1988,"children":1989},{"style":1238},[1990],{"type":50,"value":1246},{"type":44,"tag":606,"props":1992,"children":1993},{"style":1238},[1994],{"type":50,"value":1552},{"type":44,"tag":606,"props":1996,"children":1997},{"style":732},[1998],{"type":50,"value":1999}," grep",{"type":44,"tag":606,"props":2001,"children":2002},{"style":738},[2003],{"type":50,"value":2004}," -qx",{"type":44,"tag":606,"props":2006,"children":2007},{"style":738},[2008],{"type":50,"value":2009}," video\n",{"type":44,"tag":606,"props":2011,"children":2012},{"class":608,"line":700},[2013,2018,2022,2026,2031,2035,2039,2043],{"type":44,"tag":606,"props":2014,"children":2015},{"style":1748},[2016],{"type":50,"value":2017},"elif",{"type":44,"tag":606,"props":2019,"children":2020},{"style":1330},[2021],{"type":50,"value":1871},{"type":44,"tag":606,"props":2023,"children":2024},{"style":738},[2025],{"type":50,"value":1876},{"type":44,"tag":606,"props":2027,"children":2028},{"style":738},[2029],{"type":50,"value":2030}," gst-discoverer-1.0",{"type":44,"tag":606,"props":2032,"children":2033},{"style":1238},[2034],{"type":50,"value":1886},{"type":44,"tag":606,"props":2036,"children":2037},{"style":738},[2038],{"type":50,"value":1891},{"type":44,"tag":606,"props":2040,"children":2041},{"style":1238},[2042],{"type":50,"value":1896},{"type":44,"tag":606,"props":2044,"children":2045},{"style":1748},[2046],{"type":50,"value":1901},{"type":44,"tag":606,"props":2048,"children":2050},{"class":608,"line":2049},12,[2051,2056,2060,2064,2068,2072,2076,2081,2085,2090],{"type":44,"tag":606,"props":2052,"children":2053},{"style":732},[2054],{"type":50,"value":2055},"  gst-discoverer-1.0",{"type":44,"tag":606,"props":2057,"children":2058},{"style":1238},[2059],{"type":50,"value":1467},{"type":44,"tag":606,"props":2061,"children":2062},{"style":1232},[2063],{"type":50,"value":1760},{"type":44,"tag":606,"props":2065,"children":2066},{"style":1238},[2067],{"type":50,"value":1246},{"type":44,"tag":606,"props":2069,"children":2070},{"style":1238},[2071],{"type":50,"value":1552},{"type":44,"tag":606,"props":2073,"children":2074},{"style":732},[2075],{"type":50,"value":1999},{"type":44,"tag":606,"props":2077,"children":2078},{"style":738},[2079],{"type":50,"value":2080}," -qi",{"type":44,"tag":606,"props":2082,"children":2083},{"style":1238},[2084],{"type":50,"value":1623},{"type":44,"tag":606,"props":2086,"children":2087},{"style":738},[2088],{"type":50,"value":2089},"video",{"type":44,"tag":606,"props":2091,"children":2092},{"style":1238},[2093],{"type":50,"value":2094},"'\n",{"type":44,"tag":606,"props":2096,"children":2098},{"class":608,"line":2097},13,[2099],{"type":44,"tag":606,"props":2100,"children":2101},{"style":1748},[2102],{"type":50,"value":2103},"else\n",{"type":44,"tag":606,"props":2105,"children":2107},{"class":608,"line":2106},14,[2108,2113,2117,2122,2126],{"type":44,"tag":606,"props":2109,"children":2110},{"style":1330},[2111],{"type":50,"value":2112},"  echo",{"type":44,"tag":606,"props":2114,"children":2115},{"style":1238},[2116],{"type":50,"value":1467},{"type":44,"tag":606,"props":2118,"children":2119},{"style":738},[2120],{"type":50,"value":2121},"Install ffprobe or gst-discoverer-1.0 before RTSP validation.",{"type":44,"tag":606,"props":2123,"children":2124},{"style":1238},[2125],{"type":50,"value":1246},{"type":44,"tag":606,"props":2127,"children":2128},{"style":1238},[2129],{"type":50,"value":2130}," >&2\n",{"type":44,"tag":606,"props":2132,"children":2134},{"class":608,"line":2133},15,[2135,2140],{"type":44,"tag":606,"props":2136,"children":2137},{"style":1330},[2138],{"type":50,"value":2139},"  exit",{"type":44,"tag":606,"props":2141,"children":2142},{"style":1834},[2143],{"type":50,"value":2144}," 1\n",{"type":44,"tag":606,"props":2146,"children":2148},{"class":608,"line":2147},16,[2149],{"type":44,"tag":606,"props":2150,"children":2151},{"style":1748},[2152],{"type":50,"value":2153},"fi\n",{"type":44,"tag":45,"props":2155,"children":2157},{"id":2156},"quick-start-dense-captions-from-a-local-video",[2158],{"type":50,"value":2159},"Quick Start — dense captions from a local video",{"type":44,"tag":595,"props":2161,"children":2163},{"className":597,"code":2162,"language":599,"meta":600,"style":600},"# 1. Upload the video, capture its file id\nFILE_ID=$(curl -fsS -X POST \"$BASE_URL\u002Fv1\u002Ffiles\" \\\n  -H \"Authorization: Bearer $API_KEY\" \\\n  -F \"file=@\u002Fpath\u002Fto\u002Fwarehouse.mp4\" \\\n  -F \"purpose=vision\" \\\n  -F \"media_type=video\" | jq -r '.id')\n\n# 2. Generate captions + alerts (SSE stream of chunked responses)\ncurl -N -X POST \"$BASE_URL\u002Fv1\u002Fgenerate_captions\" \\\n  -H \"Authorization: Bearer $API_KEY\" \\\n  -H \"Content-Type: application\u002Fjson\" \\\n  -d \"{\n    \\\"id\\\": \\\"$FILE_ID\\\",\n    \\\"prompt\\\": \\\"Write a concise dense caption for each 10-second segment of this warehouse video.\\\",\n    \\\"model\\\": \\\"$MODEL_ID\\\",\n    \\\"chunk_duration\\\": 10,\n    \\\"stream\\\": true\n  }\"\n",[2164],{"type":44,"tag":85,"props":2165,"children":2166},{"__ignoreMap":600},[2167,2175,2227,2255,2280,2304,2354,2361,2369,2410,2437,2461,2478,2510,2547,2576,2597,2619],{"type":44,"tag":606,"props":2168,"children":2169},{"class":608,"line":609},[2170],{"type":44,"tag":606,"props":2171,"children":2172},{"style":613},[2173],{"type":50,"value":2174},"# 1. Upload the video, capture its file id\n",{"type":44,"tag":606,"props":2176,"children":2177},{"class":608,"line":619},[2178,2183,2188,2192,2196,2201,2206,2210,2214,2219,2223],{"type":44,"tag":606,"props":2179,"children":2180},{"style":1232},[2181],{"type":50,"value":2182},"FILE_ID",{"type":44,"tag":606,"props":2184,"children":2185},{"style":1238},[2186],{"type":50,"value":2187},"=$(",{"type":44,"tag":606,"props":2189,"children":2190},{"style":732},[2191],{"type":50,"value":109},{"type":44,"tag":606,"props":2193,"children":2194},{"style":738},[2195],{"type":50,"value":1462},{"type":44,"tag":606,"props":2197,"children":2198},{"style":738},[2199],{"type":50,"value":2200}," -X",{"type":44,"tag":606,"props":2202,"children":2203},{"style":738},[2204],{"type":50,"value":2205}," POST",{"type":44,"tag":606,"props":2207,"children":2208},{"style":1238},[2209],{"type":50,"value":1467},{"type":44,"tag":606,"props":2211,"children":2212},{"style":1232},[2213],{"type":50,"value":138},{"type":44,"tag":606,"props":2215,"children":2216},{"style":738},[2217],{"type":50,"value":2218},"\u002Fv1\u002Ffiles",{"type":44,"tag":606,"props":2220,"children":2221},{"style":1238},[2222],{"type":50,"value":1246},{"type":44,"tag":606,"props":2224,"children":2225},{"style":1232},[2226],{"type":50,"value":1933},{"type":44,"tag":606,"props":2228,"children":2229},{"class":608,"line":628},[2230,2235,2239,2243,2247,2251],{"type":44,"tag":606,"props":2231,"children":2232},{"style":738},[2233],{"type":50,"value":2234},"  -H",{"type":44,"tag":606,"props":2236,"children":2237},{"style":1238},[2238],{"type":50,"value":1467},{"type":44,"tag":606,"props":2240,"children":2241},{"style":738},[2242],{"type":50,"value":1538},{"type":44,"tag":606,"props":2244,"children":2245},{"style":1232},[2246],{"type":50,"value":1543},{"type":44,"tag":606,"props":2248,"children":2249},{"style":1238},[2250],{"type":50,"value":1246},{"type":44,"tag":606,"props":2252,"children":2253},{"style":1232},[2254],{"type":50,"value":1933},{"type":44,"tag":606,"props":2256,"children":2257},{"class":608,"line":637},[2258,2263,2267,2272,2276],{"type":44,"tag":606,"props":2259,"children":2260},{"style":738},[2261],{"type":50,"value":2262},"  -F",{"type":44,"tag":606,"props":2264,"children":2265},{"style":1238},[2266],{"type":50,"value":1467},{"type":44,"tag":606,"props":2268,"children":2269},{"style":738},[2270],{"type":50,"value":2271},"file=@\u002Fpath\u002Fto\u002Fwarehouse.mp4",{"type":44,"tag":606,"props":2273,"children":2274},{"style":1238},[2275],{"type":50,"value":1246},{"type":44,"tag":606,"props":2277,"children":2278},{"style":1232},[2279],{"type":50,"value":1933},{"type":44,"tag":606,"props":2281,"children":2282},{"class":608,"line":646},[2283,2287,2291,2296,2300],{"type":44,"tag":606,"props":2284,"children":2285},{"style":738},[2286],{"type":50,"value":2262},{"type":44,"tag":606,"props":2288,"children":2289},{"style":1238},[2290],{"type":50,"value":1467},{"type":44,"tag":606,"props":2292,"children":2293},{"style":738},[2294],{"type":50,"value":2295},"purpose=vision",{"type":44,"tag":606,"props":2297,"children":2298},{"style":1238},[2299],{"type":50,"value":1246},{"type":44,"tag":606,"props":2301,"children":2302},{"style":1232},[2303],{"type":50,"value":1933},{"type":44,"tag":606,"props":2305,"children":2306},{"class":608,"line":655},[2307,2311,2315,2320,2324,2328,2332,2336,2340,2345,2349],{"type":44,"tag":606,"props":2308,"children":2309},{"style":738},[2310],{"type":50,"value":2262},{"type":44,"tag":606,"props":2312,"children":2313},{"style":1238},[2314],{"type":50,"value":1467},{"type":44,"tag":606,"props":2316,"children":2317},{"style":738},[2318],{"type":50,"value":2319},"media_type=video",{"type":44,"tag":606,"props":2321,"children":2322},{"style":1238},[2323],{"type":50,"value":1246},{"type":44,"tag":606,"props":2325,"children":2326},{"style":1238},[2327],{"type":50,"value":1552},{"type":44,"tag":606,"props":2329,"children":2330},{"style":732},[2331],{"type":50,"value":1557},{"type":44,"tag":606,"props":2333,"children":2334},{"style":738},[2335],{"type":50,"value":1618},{"type":44,"tag":606,"props":2337,"children":2338},{"style":1238},[2339],{"type":50,"value":1623},{"type":44,"tag":606,"props":2341,"children":2342},{"style":738},[2343],{"type":50,"value":2344},".id",{"type":44,"tag":606,"props":2346,"children":2347},{"style":1238},[2348],{"type":50,"value":1567},{"type":44,"tag":606,"props":2350,"children":2351},{"style":1238},[2352],{"type":50,"value":2353},")\n",{"type":44,"tag":606,"props":2355,"children":2356},{"class":608,"line":664},[2357],{"type":44,"tag":606,"props":2358,"children":2359},{"emptyLinePlaceholder":1855},[2360],{"type":50,"value":1858},{"type":44,"tag":606,"props":2362,"children":2363},{"class":608,"line":673},[2364],{"type":44,"tag":606,"props":2365,"children":2366},{"style":613},[2367],{"type":50,"value":2368},"# 2. Generate captions + alerts (SSE stream of chunked responses)\n",{"type":44,"tag":606,"props":2370,"children":2371},{"class":608,"line":682},[2372,2376,2381,2385,2389,2393,2397,2402,2406],{"type":44,"tag":606,"props":2373,"children":2374},{"style":732},[2375],{"type":50,"value":109},{"type":44,"tag":606,"props":2377,"children":2378},{"style":738},[2379],{"type":50,"value":2380}," -N",{"type":44,"tag":606,"props":2382,"children":2383},{"style":738},[2384],{"type":50,"value":2200},{"type":44,"tag":606,"props":2386,"children":2387},{"style":738},[2388],{"type":50,"value":2205},{"type":44,"tag":606,"props":2390,"children":2391},{"style":1238},[2392],{"type":50,"value":1467},{"type":44,"tag":606,"props":2394,"children":2395},{"style":1232},[2396],{"type":50,"value":138},{"type":44,"tag":606,"props":2398,"children":2399},{"style":738},[2400],{"type":50,"value":2401},"\u002Fv1\u002Fgenerate_captions",{"type":44,"tag":606,"props":2403,"children":2404},{"style":1238},[2405],{"type":50,"value":1246},{"type":44,"tag":606,"props":2407,"children":2408},{"style":1232},[2409],{"type":50,"value":1933},{"type":44,"tag":606,"props":2411,"children":2412},{"class":608,"line":691},[2413,2417,2421,2425,2429,2433],{"type":44,"tag":606,"props":2414,"children":2415},{"style":738},[2416],{"type":50,"value":2234},{"type":44,"tag":606,"props":2418,"children":2419},{"style":1238},[2420],{"type":50,"value":1467},{"type":44,"tag":606,"props":2422,"children":2423},{"style":738},[2424],{"type":50,"value":1538},{"type":44,"tag":606,"props":2426,"children":2427},{"style":1232},[2428],{"type":50,"value":1543},{"type":44,"tag":606,"props":2430,"children":2431},{"style":1238},[2432],{"type":50,"value":1246},{"type":44,"tag":606,"props":2434,"children":2435},{"style":1232},[2436],{"type":50,"value":1933},{"type":44,"tag":606,"props":2438,"children":2439},{"class":608,"line":700},[2440,2444,2448,2453,2457],{"type":44,"tag":606,"props":2441,"children":2442},{"style":738},[2443],{"type":50,"value":2234},{"type":44,"tag":606,"props":2445,"children":2446},{"style":1238},[2447],{"type":50,"value":1467},{"type":44,"tag":606,"props":2449,"children":2450},{"style":738},[2451],{"type":50,"value":2452},"Content-Type: application\u002Fjson",{"type":44,"tag":606,"props":2454,"children":2455},{"style":1238},[2456],{"type":50,"value":1246},{"type":44,"tag":606,"props":2458,"children":2459},{"style":1232},[2460],{"type":50,"value":1933},{"type":44,"tag":606,"props":2462,"children":2463},{"class":608,"line":2049},[2464,2469,2473],{"type":44,"tag":606,"props":2465,"children":2466},{"style":738},[2467],{"type":50,"value":2468},"  -d",{"type":44,"tag":606,"props":2470,"children":2471},{"style":1238},[2472],{"type":50,"value":1467},{"type":44,"tag":606,"props":2474,"children":2475},{"style":738},[2476],{"type":50,"value":2477},"{\n",{"type":44,"tag":606,"props":2479,"children":2480},{"class":608,"line":2097},[2481,2486,2491,2496,2500,2505],{"type":44,"tag":606,"props":2482,"children":2483},{"style":1232},[2484],{"type":50,"value":2485},"    \\\"",{"type":44,"tag":606,"props":2487,"children":2488},{"style":738},[2489],{"type":50,"value":2490},"id",{"type":44,"tag":606,"props":2492,"children":2493},{"style":1232},[2494],{"type":50,"value":2495},"\\\"",{"type":44,"tag":606,"props":2497,"children":2498},{"style":738},[2499],{"type":50,"value":429},{"type":44,"tag":606,"props":2501,"children":2502},{"style":1232},[2503],{"type":50,"value":2504},"\\\"$FILE_ID\\\"",{"type":44,"tag":606,"props":2506,"children":2507},{"style":738},[2508],{"type":50,"value":2509},",\n",{"type":44,"tag":606,"props":2511,"children":2512},{"class":608,"line":2106},[2513,2517,2522,2526,2530,2534,2539,2543],{"type":44,"tag":606,"props":2514,"children":2515},{"style":1232},[2516],{"type":50,"value":2485},{"type":44,"tag":606,"props":2518,"children":2519},{"style":738},[2520],{"type":50,"value":2521},"prompt",{"type":44,"tag":606,"props":2523,"children":2524},{"style":1232},[2525],{"type":50,"value":2495},{"type":44,"tag":606,"props":2527,"children":2528},{"style":738},[2529],{"type":50,"value":429},{"type":44,"tag":606,"props":2531,"children":2532},{"style":1232},[2533],{"type":50,"value":2495},{"type":44,"tag":606,"props":2535,"children":2536},{"style":738},[2537],{"type":50,"value":2538},"Write a concise dense caption for each 10-second segment of this warehouse video.",{"type":44,"tag":606,"props":2540,"children":2541},{"style":1232},[2542],{"type":50,"value":2495},{"type":44,"tag":606,"props":2544,"children":2545},{"style":738},[2546],{"type":50,"value":2509},{"type":44,"tag":606,"props":2548,"children":2549},{"class":608,"line":2133},[2550,2554,2559,2563,2567,2572],{"type":44,"tag":606,"props":2551,"children":2552},{"style":1232},[2553],{"type":50,"value":2485},{"type":44,"tag":606,"props":2555,"children":2556},{"style":738},[2557],{"type":50,"value":2558},"model",{"type":44,"tag":606,"props":2560,"children":2561},{"style":1232},[2562],{"type":50,"value":2495},{"type":44,"tag":606,"props":2564,"children":2565},{"style":738},[2566],{"type":50,"value":429},{"type":44,"tag":606,"props":2568,"children":2569},{"style":1232},[2570],{"type":50,"value":2571},"\\\"$MODEL_ID\\\"",{"type":44,"tag":606,"props":2573,"children":2574},{"style":738},[2575],{"type":50,"value":2509},{"type":44,"tag":606,"props":2577,"children":2578},{"class":608,"line":2147},[2579,2583,2588,2592],{"type":44,"tag":606,"props":2580,"children":2581},{"style":1232},[2582],{"type":50,"value":2485},{"type":44,"tag":606,"props":2584,"children":2585},{"style":738},[2586],{"type":50,"value":2587},"chunk_duration",{"type":44,"tag":606,"props":2589,"children":2590},{"style":1232},[2591],{"type":50,"value":2495},{"type":44,"tag":606,"props":2593,"children":2594},{"style":738},[2595],{"type":50,"value":2596},": 10,\n",{"type":44,"tag":606,"props":2598,"children":2600},{"class":608,"line":2599},17,[2601,2605,2610,2614],{"type":44,"tag":606,"props":2602,"children":2603},{"style":1232},[2604],{"type":50,"value":2485},{"type":44,"tag":606,"props":2606,"children":2607},{"style":738},[2608],{"type":50,"value":2609},"stream",{"type":44,"tag":606,"props":2611,"children":2612},{"style":1232},[2613],{"type":50,"value":2495},{"type":44,"tag":606,"props":2615,"children":2616},{"style":738},[2617],{"type":50,"value":2618},": true\n",{"type":44,"tag":606,"props":2620,"children":2622},{"class":608,"line":2621},18,[2623,2628],{"type":44,"tag":606,"props":2624,"children":2625},{"style":738},[2626],{"type":50,"value":2627},"  }",{"type":44,"tag":606,"props":2629,"children":2630},{"style":1238},[2631],{"type":50,"value":1481},{"type":44,"tag":45,"props":2633,"children":2635},{"id":2634},"api-surface",[2636],{"type":50,"value":2637},"API Surface",{"type":44,"tag":53,"props":2639,"children":2640},{},[2641],{"type":50,"value":2642},"Use the live OpenAPI as the source of truth before calling optional endpoints:",{"type":44,"tag":595,"props":2644,"children":2646},{"className":597,"code":2645,"language":599,"meta":600,"style":600},"curl -fsS \"$BASE_URL\u002Fopenapi.json\" | jq -r '.paths | keys[]' | sort\n",[2647],{"type":44,"tag":85,"props":2648,"children":2649},{"__ignoreMap":600},[2650],{"type":44,"tag":606,"props":2651,"children":2652},{"class":608,"line":609},[2653,2657,2661,2665,2669,2673,2677,2681,2685,2689,2693,2697,2701,2705],{"type":44,"tag":606,"props":2654,"children":2655},{"style":732},[2656],{"type":50,"value":109},{"type":44,"tag":606,"props":2658,"children":2659},{"style":738},[2660],{"type":50,"value":1462},{"type":44,"tag":606,"props":2662,"children":2663},{"style":1238},[2664],{"type":50,"value":1467},{"type":44,"tag":606,"props":2666,"children":2667},{"style":1232},[2668],{"type":50,"value":138},{"type":44,"tag":606,"props":2670,"children":2671},{"style":738},[2672],{"type":50,"value":1601},{"type":44,"tag":606,"props":2674,"children":2675},{"style":1238},[2676],{"type":50,"value":1246},{"type":44,"tag":606,"props":2678,"children":2679},{"style":1238},[2680],{"type":50,"value":1552},{"type":44,"tag":606,"props":2682,"children":2683},{"style":732},[2684],{"type":50,"value":1557},{"type":44,"tag":606,"props":2686,"children":2687},{"style":738},[2688],{"type":50,"value":1618},{"type":44,"tag":606,"props":2690,"children":2691},{"style":1238},[2692],{"type":50,"value":1623},{"type":44,"tag":606,"props":2694,"children":2695},{"style":738},[2696],{"type":50,"value":1628},{"type":44,"tag":606,"props":2698,"children":2699},{"style":1238},[2700],{"type":50,"value":1567},{"type":44,"tag":606,"props":2702,"children":2703},{"style":1238},[2704],{"type":50,"value":1552},{"type":44,"tag":606,"props":2706,"children":2707},{"style":732},[2708],{"type":50,"value":1641},{"type":44,"tag":53,"props":2710,"children":2711},{},[2712],{"type":50,"value":2713},"Core paths for VSS 3.2 are:",{"type":44,"tag":70,"props":2715,"children":2716},{},[2717,2735,2768,2801,2820],{"type":44,"tag":74,"props":2718,"children":2719},{},[2720,2726,2728,2733],{"type":44,"tag":85,"props":2721,"children":2723},{"className":2722},[],[2724],{"type":50,"value":2725},"POST \u002Fv1\u002Ffiles",{"type":50,"value":2727}," for multipart media upload; pass the returned file ",{"type":44,"tag":85,"props":2729,"children":2731},{"className":2730},[],[2732],{"type":50,"value":2490},{"type":50,"value":2734}," into\ncaption generation and delete the file when finished.",{"type":44,"tag":74,"props":2736,"children":2737},{},[2738,2744,2746,2752,2754,2759,2761,2766],{"type":44,"tag":85,"props":2739,"children":2741},{"className":2740},[],[2742],{"type":50,"value":2743},"POST \u002Fv1\u002Fgenerate_captions",{"type":50,"value":2745}," for file or stream captioning. Use the exact\nmodel id returned by ",{"type":44,"tag":85,"props":2747,"children":2749},{"className":2748},[],[2750],{"type":50,"value":2751},"GET \u002Fv1\u002Fmodels",{"type":50,"value":2753},"; aliases such as ",{"type":44,"tag":85,"props":2755,"children":2757},{"className":2756},[],[2758],{"type":50,"value":489},{"type":50,"value":2760}," or\n",{"type":44,"tag":85,"props":2762,"children":2764},{"className":2763},[],[2765],{"type":50,"value":496},{"type":50,"value":2767}," are backend selectors, not request model ids.",{"type":44,"tag":74,"props":2769,"children":2770},{},[2771,2777,2778,2784,2786,2792,2794,2800],{"type":44,"tag":85,"props":2772,"children":2774},{"className":2773},[],[2775],{"type":50,"value":2776},"POST \u002Fv1\u002Fstreams\u002Fadd",{"type":50,"value":111},{"type":44,"tag":85,"props":2779,"children":2781},{"className":2780},[],[2782],{"type":50,"value":2783},"GET \u002Fv1\u002Fstreams\u002Fget-stream-info",{"type":50,"value":2785},", and\n",{"type":44,"tag":85,"props":2787,"children":2789},{"className":2788},[],[2790],{"type":50,"value":2791},"DELETE \u002Fv1\u002Fstreams\u002Fdelete\u002F{stream_id}",{"type":50,"value":2793}," for RTSP lifecycle. Parse stream ids\nfrom ",{"type":44,"tag":85,"props":2795,"children":2797},{"className":2796},[],[2798],{"type":50,"value":2799},"results[0].id",{"type":50,"value":140},{"type":44,"tag":74,"props":2802,"children":2803},{},[2804,2810,2812,2818],{"type":44,"tag":85,"props":2805,"children":2807},{"className":2806},[],[2808],{"type":50,"value":2809},"POST \u002Fv1\u002Fchat\u002Fcompletions",{"type":50,"value":2811}," for OpenAI-compatible text and multimodal calls.\nCurrent 26.05 builds return HTTP 400 for text-only ",{"type":44,"tag":85,"props":2813,"children":2815},{"className":2814},[],[2816],{"type":50,"value":2817},"\u002Fv1\u002Fcompletions",{"type":50,"value":2819},"; treat\nthat as expected when validating legacy behavior.",{"type":44,"tag":74,"props":2821,"children":2822},{},[2823,2829,2830,2835,2836,2842,2843,2849,2851,2857],{"type":44,"tag":85,"props":2824,"children":2826},{"className":2825},[],[2827],{"type":50,"value":2828},"GET \u002Fv1\u002Fhealth\u002Fready",{"type":50,"value":111},{"type":44,"tag":85,"props":2831,"children":2833},{"className":2832},[],[2834],{"type":50,"value":1520},{"type":50,"value":111},{"type":44,"tag":85,"props":2837,"children":2839},{"className":2838},[],[2840],{"type":50,"value":2841},"\u002Fv1\u002Fassets\u002Fstats",{"type":50,"value":307},{"type":44,"tag":85,"props":2844,"children":2846},{"className":2845},[],[2847],{"type":50,"value":2848},"\u002Fv1\u002Fmetrics",{"type":50,"value":2850},"\nfor service probes. Do not assume ",{"type":44,"tag":85,"props":2852,"children":2854},{"className":2853},[],[2855],{"type":50,"value":2856},"\u002Fv1\u002Flicense",{"type":50,"value":2858}," exists unless OpenAPI lists it.",{"type":44,"tag":53,"props":2860,"children":2861},{},[2862,2864,2873],{"type":50,"value":2863},"Detailed endpoint schemas, response shapes, CV-style singular stream endpoints,\nand 26.05 compatibility notes live in\n",{"type":44,"tag":508,"props":2865,"children":2867},{"href":2866},"references\u002Fapi-surface-26.05.md",[2868],{"type":44,"tag":85,"props":2869,"children":2871},{"className":2870},[],[2872],{"type":50,"value":2866},{"type":50,"value":140},{"type":44,"tag":45,"props":2875,"children":2877},{"id":2876},"common-workflows",[2878],{"type":50,"value":2879},"Common Workflows",{"type":44,"tag":70,"props":2881,"children":2882},{},[2883,2910,2935,2958],{"type":44,"tag":74,"props":2884,"children":2885},{},[2886,2888,2893,2895,2900,2902,2908],{"type":50,"value":2887},"Stored file captioning: upload with ",{"type":44,"tag":85,"props":2889,"children":2891},{"className":2890},[],[2892],{"type":50,"value":2725},{"type":50,"value":2894},", call\n",{"type":44,"tag":85,"props":2896,"children":2898},{"className":2897},[],[2899],{"type":50,"value":2401},{"type":50,"value":2901}," with the returned file id, use ",{"type":44,"tag":85,"props":2903,"children":2905},{"className":2904},[],[2906],{"type":50,"value":2907},"stream=true",{"type":50,"value":2909}," for SSE,\nthen delete the file to release storage.",{"type":44,"tag":74,"props":2911,"children":2912},{},[2913,2915,2920,2922,2926,2928,2933],{"type":50,"value":2914},"RTSP live captioning: when the caller provides ",{"type":44,"tag":85,"props":2916,"children":2918},{"className":2917},[],[2919],{"type":50,"value":1658},{"type":50,"value":2921},", use that\nexact URL and run the ",{"type":44,"tag":350,"props":2923,"children":2924},{},[2925],{"type":50,"value":1647},{"type":50,"value":2927}," before registration. Do not\nderive a replacement stream from NvStreamer or VIOS when ",{"type":44,"tag":85,"props":2929,"children":2931},{"className":2930},[],[2932],{"type":50,"value":1658},{"type":50,"value":2934}," is\nempty; fail fast instead. Require an actual video stream\u002Fcaps entry before\nregistration; add the stream, caption it, then unregister it.",{"type":44,"tag":74,"props":2936,"children":2937},{},[2938,2940,2946,2948,2957],{"type":50,"value":2939},"Alert prompts: include a deterministic ",{"type":44,"tag":85,"props":2941,"children":2943},{"className":2942},[],[2944],{"type":50,"value":2945},"Anomaly Detected: Yes\u002FNo",{"type":50,"value":2947}," line.\nKafka publication is server-side config, additive to HTTP responses, and\ndocumented in ",{"type":44,"tag":508,"props":2949,"children":2951},{"href":2950},"references\u002Fkafka-workflows.md",[2952],{"type":44,"tag":85,"props":2953,"children":2955},{"className":2954},[],[2956],{"type":50,"value":2950},{"type":50,"value":140},{"type":44,"tag":74,"props":2959,"children":2960},{},[2961,2963,2969,2971,2977,2979,2984],{"type":50,"value":2962},"Kafka validation: trust the live ",{"type":44,"tag":85,"props":2964,"children":2966},{"className":2965},[],[2967],{"type":50,"value":2968},"vss-rtvi-vlm",{"type":50,"value":2970}," environment for topic names.\nIn a full VSS alerts real-time profile, use the existing VSS Kafka container\n",{"type":44,"tag":85,"props":2972,"children":2974},{"className":2973},[],[2975],{"type":50,"value":2976},"mdx-kafka",{"type":50,"value":2978}," for CLI checks and final incident-consumer commands. For\nstandalone validation, use a broker that advertises ",{"type":44,"tag":85,"props":2980,"children":2982},{"className":2981},[],[2983],{"type":50,"value":1068},{"type":50,"value":2985},"; never\nstop or replace a pre-existing broker without user confirmation.",{"type":44,"tag":45,"props":2987,"children":2989},{"id":2988},"error-reference",[2990],{"type":50,"value":2991},"Error Reference",{"type":44,"tag":53,"props":2993,"children":2994},{},[2995,2997,3003],{"type":50,"value":2996},"Common causes: 400 for invalid request shape or model id, 401\u002F403 for missing\nor wrong bearer token, 404 for deleted files\u002Fstreams or unsupported endpoints,\n413 for oversized uploads, 422 for schema validation, 429 for too much\nconcurrency, 500 for inference\u002Fruntime failures, and 503 while startup is still\nin progress. Inspect ",{"type":44,"tag":85,"props":2998,"children":3000},{"className":2999},[],[3001],{"type":50,"value":3002},"docker logs vss-rtvi-vlm",{"type":50,"value":3004}," for service-side failures.",{"type":44,"tag":3006,"props":3007,"children":3008},"style",{},[3009],{"type":50,"value":3010},"html .light .shiki span {color: var(--shiki-light);background: var(--shiki-light-bg);font-style: var(--shiki-light-font-style);font-weight: var(--shiki-light-font-weight);text-decoration: var(--shiki-light-text-decoration);}html.light .shiki span {color: var(--shiki-light);background: var(--shiki-light-bg);font-style: var(--shiki-light-font-style);font-weight: var(--shiki-light-font-weight);text-decoration: var(--shiki-light-text-decoration);}html .default .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}html .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}html .dark .shiki span {color: var(--shiki-dark);background: var(--shiki-dark-bg);font-style: var(--shiki-dark-font-style);font-weight: var(--shiki-dark-font-weight);text-decoration: var(--shiki-dark-text-decoration);}html.dark .shiki span {color: var(--shiki-dark);background: var(--shiki-dark-bg);font-style: var(--shiki-dark-font-style);font-weight: var(--shiki-dark-font-weight);text-decoration: var(--shiki-dark-text-decoration);}",{"items":3012,"total":3112},[3013,3030,3042,3056,3068,3084,3098],{"slug":3014,"name":3014,"fn":3015,"description":3016,"org":3017,"tags":3018,"stars":23,"repoUrl":24,"updatedAt":3029},"accelerated-computing-cudf","accelerate data processing with cuDF","Official NVIDIA-authored guidance for NVIDIA cuDF GPU DataFrames, pandas acceleration, dask-cuDF, ETL, joins, groupby, CSV\u002FParquet I\u002FO, nullable semantics, and multi-GPU DataFrame workloads.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[3019,3022,3025,3026],{"name":3020,"slug":3021,"type":15},"Data Analysis","data-analysis",{"name":3023,"slug":3024,"type":15},"Data Engineering","data-engineering",{"name":9,"slug":8,"type":15},{"name":3027,"slug":3028,"type":15},"Performance","performance","2026-07-14T05:28:43.176466",{"slug":3031,"name":3031,"fn":3032,"description":3033,"org":3034,"tags":3035,"stars":23,"repoUrl":24,"updatedAt":3041},"aiq-deploy","deploy and manage NVIDIA AI-Q infrastructure","Use when asked to install, deploy, run, validate, troubleshoot, or stop NVIDIA AI-Q Blueprint infrastructure.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[3036,3037,3040],{"name":13,"slug":14,"type":15},{"name":3038,"slug":3039,"type":15},"Infrastructure","infrastructure",{"name":9,"slug":8,"type":15},"2026-07-14T05:29:06.667109",{"slug":3043,"name":3043,"fn":3044,"description":3045,"org":3046,"tags":3047,"stars":23,"repoUrl":24,"updatedAt":3055},"aiq-research","conduct deep research with AI-Q","Use when asked to run deep research or AI-Q research through a reachable NVIDIA AI-Q Blueprint backend.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[3048,3051,3052],{"name":3049,"slug":3050,"type":15},"Agents","agents",{"name":9,"slug":8,"type":15},{"name":3053,"slug":3054,"type":15},"Research","research","2026-07-14T05:28:06.816956",{"slug":3057,"name":3057,"fn":3058,"description":3059,"org":3060,"tags":3061,"stars":23,"repoUrl":24,"updatedAt":3067},"amc-run-sample-calibration","run AMC sample dataset calibration","Run end-to-end calibration on the shipped sample dataset (sdg_08_2_sample_data_010926.zip) against a running AMC microservice. Use when user says 'test sample dataset', 'run sample calibration', 'verify AMC install', or 'launch and test'.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[3062,3063,3064],{"name":3020,"slug":3021,"type":15},{"name":9,"slug":8,"type":15},{"name":3065,"slug":3066,"type":15},"Testing","testing","2026-07-17T05:29:03.913266",{"slug":3069,"name":3069,"fn":3070,"description":3071,"org":3072,"tags":3073,"stars":23,"repoUrl":24,"updatedAt":3083},"amc-run-video-calibration","calibrate video datasets with AutoMagicCalib","Calibrate a new dataset from pre-recorded video files via the AutoMagicCalib REST API. Use when user has local MP4s and says 'calibrate my videos', 'run AMC on these videos', or similar. For RTSP\u002Flive streams, use amc-run-rtsp-calibration instead.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[3074,3077,3080,3081],{"name":3075,"slug":3076,"type":15},"Automation","automation",{"name":3078,"slug":3079,"type":15},"Imaging","imaging",{"name":9,"slug":8,"type":15},{"name":3082,"slug":2089,"type":15},"Video","2026-07-17T05:28:53.905004",{"slug":3085,"name":3085,"fn":3086,"description":3087,"org":3088,"tags":3089,"stars":23,"repoUrl":24,"updatedAt":3097},"amc-setup-calibration-stack","deploy AutoMagicCalib microservice with Docker","Launch AutoMagicCalib microservice and web UI from NGC release images via Docker Compose. Use when user says 'deploy auto calibration', 'launch auto calibration', 'launch AMC', 'start MS+UI', or 'set up auto-magic-calib'. Requires NGC API key.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[3090,3091,3093,3094],{"name":13,"slug":14,"type":15},{"name":3092,"slug":767,"type":15},"Docker",{"name":9,"slug":8,"type":15},{"name":3095,"slug":3096,"type":15},"Operations","operations","2026-07-17T05:28:56.913999",{"slug":3099,"name":3099,"fn":3100,"description":3101,"org":3102,"tags":3103,"stars":23,"repoUrl":24,"updatedAt":3111},"cudaq-guide","develop quantum applications with CUDA-Q","CUDA-Q onboarding guide for installation, test programs, GPU simulation, QPU hardware, and quantum applications.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[3104,3105,3108],{"name":9,"slug":8,"type":15},{"name":3106,"slug":3107,"type":15},"Quantum Computing","quantum-computing",{"name":3109,"slug":3110,"type":15},"Simulation","simulation","2026-07-14T05:26:58.898253",305,{"items":3114,"total":3265},[3115,3133,3149,3160,3172,3186,3199,3213,3224,3233,3247,3256],{"slug":3116,"name":3116,"fn":3117,"description":3118,"org":3119,"tags":3120,"stars":3130,"repoUrl":3131,"updatedAt":3132},"nemoclaw-user-guide","retrieve NemoClaw documentation and configuration","Guides human users' AI agents to the NemoClaw docs MCP server and canonical Fern documentation in Markdown form. Use when users ask how to install, configure, operate, troubleshoot, secure, or learn NemoClaw with an AI coding assistant. Trigger keywords - nemoclaw docs, use nemoclaw with ai agent, nemoclaw mcp docs, nemoclaw install help, nemoclaw quickstart, nemoclaw markdown docs, llms.txt, agent skills.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[3121,3124,3127],{"name":3122,"slug":3123,"type":15},"Documentation","documentation",{"name":3125,"slug":3126,"type":15},"MCP","mcp",{"name":3128,"slug":3129,"type":15},"Search","search",21777,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FNemoClaw","2026-07-20T06:00:01.461044",{"slug":3134,"name":3134,"fn":3135,"description":3136,"org":3137,"tags":3138,"stars":3146,"repoUrl":3147,"updatedAt":3148},"mcore-build-and-dependency","manage Megatron-LM development environments","Container-based dev environment setup and dependency management for Megatron-LM. Covers acquiring and launching the CI container, uv package management, and updating uv.lock.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[3139,3142,3143],{"name":3140,"slug":3141,"type":15},"Containers","containers",{"name":13,"slug":14,"type":15},{"name":3144,"slug":3145,"type":15},"Python","python",17049,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FMegatron-LM","2026-07-27T06:06:11.249662",{"slug":3150,"name":3150,"fn":3151,"description":3152,"org":3153,"tags":3154,"stars":3146,"repoUrl":3147,"updatedAt":3159},"mcore-bump-base-image","update NVIDIA PyTorch base images","Bump the NVIDIA PyTorch base image (`nvcr.io\u002Fnvidia\u002Fpytorch:YY.MM-py3`) used by Megatron-LM CI. Covers the two pin sites (GitHub CI in `docker\u002F.ngc_version.dev` and GitLab CI in `.gitlab\u002Fstages\u002F01.build.yml`), the post-bump CI loop (re-run functional tests, refresh golden values, mark broken tests), and the gotchas that bit PRs",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[3155,3158],{"name":3156,"slug":3157,"type":15},"CI\u002FCD","ci-cd",{"name":13,"slug":14,"type":15},"2026-07-14T05:25:59.97109",{"slug":3161,"name":3161,"fn":3162,"description":3163,"org":3164,"tags":3165,"stars":3146,"repoUrl":3147,"updatedAt":3171},"mcore-cicd","manage CI\u002FCD pipelines for Megatron-LM","CI\u002FCD reference for Megatron-LM. Covers CI pipeline structure, PR scope labels, triggering internal GitLab CI (which force-pushes the current branch to a pull-request\u002FBRANCH ref — always dry-run and verify the destination first; never run against shared or protected branches), and CI failure investigation.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[3166,3167,3168],{"name":3156,"slug":3157,"type":15},{"name":13,"slug":14,"type":15},{"name":3169,"slug":3170,"type":15},"GitHub","github","2026-07-27T06:06:12.278222",{"slug":3173,"name":3173,"fn":3174,"description":3175,"org":3176,"tags":3177,"stars":3146,"repoUrl":3147,"updatedAt":3185},"mcore-create-issue","investigate CI failures and create issues","Investigate a failing GitHub Actions run or job and create a GitHub issue for the failure.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[3178,3181,3182],{"name":3179,"slug":3180,"type":15},"Debugging","debugging",{"name":3169,"slug":3170,"type":15},{"name":3183,"slug":3184,"type":15},"Triage","triage","2026-07-14T05:25:57.442089",{"slug":3187,"name":3187,"fn":3188,"description":3189,"org":3190,"tags":3191,"stars":3146,"repoUrl":3147,"updatedAt":3198},"mcore-linting-and-formatting","lint and format Megatron-LM code","Linting and formatting for Megatron-LM. Covers running autoformat.sh, tools (ruff, black, isort, pylint, mypy), and code style rules.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[3192,3195],{"name":3193,"slug":3194,"type":15},"Best Practices","best-practices",{"name":3196,"slug":3197,"type":15},"Code Analysis","code-analysis","2026-07-14T05:25:56.18433",{"slug":3200,"name":3200,"fn":3201,"description":3202,"org":3203,"tags":3204,"stars":3146,"repoUrl":3147,"updatedAt":3212},"mcore-migrate-gpt-to-hybrid","migrate Megatron-LM models to HybridModel","Migration guide for moving Megatron Core GPTModel checkpoints, model providers, training commands, and layer mappings to HybridModel.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[3205,3208,3211],{"name":3206,"slug":3207,"type":15},"Machine Learning","machine-learning",{"name":3209,"slug":3210,"type":15},"Migration","migration",{"name":9,"slug":8,"type":15},"2026-07-17T06:07:11.777011",{"slug":3214,"name":3214,"fn":3215,"description":3216,"org":3217,"tags":3218,"stars":3146,"repoUrl":3147,"updatedAt":3223},"mcore-onboard-gb200-1node-tests","onboard functional tests for GB200","Onboard 1-node GitHub MR functional tests for GB200 from existing mr-scoped 2-node tests.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[3219,3222],{"name":3220,"slug":3221,"type":15},"QA","qa",{"name":3065,"slug":3066,"type":15},"2026-07-14T05:25:53.673039",{"slug":3225,"name":3225,"fn":3226,"description":3227,"org":3228,"tags":3229,"stars":3146,"repoUrl":3147,"updatedAt":3232},"mcore-run-on-slurm","launch distributed training jobs on SLURM","How to launch distributed Megatron-LM training jobs on a SLURM cluster. Covers a minimal sbatch skeleton, environment-variable setup for torch.distributed.run, CUDA_DEVICE_MAX_CONNECTIONS rules across hardware and parallelism modes, container conventions, monitoring, and per-rank failure diagnosis.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[3230,3231],{"name":13,"slug":14,"type":15},{"name":3038,"slug":3039,"type":15},"2026-07-14T05:25:49.362534",{"slug":3234,"name":3234,"fn":3235,"description":3236,"org":3237,"tags":3238,"stars":3146,"repoUrl":3147,"updatedAt":3246},"mcore-split-pr","split pull requests to reduce review load","Split a PR into multiple PRs to reduce the number of required CODEOWNERS reviewer groups.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[3239,3242,3243],{"name":3240,"slug":3241,"type":15},"Code Review","code-review",{"name":3169,"slug":3170,"type":15},{"name":3244,"slug":3245,"type":15},"Pull Requests","pull-requests","2026-07-14T05:26:01.226578",{"slug":3248,"name":3248,"fn":3249,"description":3250,"org":3251,"tags":3252,"stars":3146,"repoUrl":3147,"updatedAt":3255},"mcore-testing","run and manage Megatron-LM tests","Test system for Megatron-LM. Covers test layout, recipe YAML structure, adding and running unit and functional tests, golden values, marker filters, and CI parity.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[3253,3254],{"name":3220,"slug":3221,"type":15},{"name":3065,"slug":3066,"type":15},"2026-07-14T05:25:54.928983",{"slug":3257,"name":3257,"fn":3258,"description":3259,"org":3260,"tags":3261,"stars":3146,"repoUrl":3147,"updatedAt":3264},"nightly-sync","manage nightly main-to-dev sync workflows","Domain knowledge for the nightly main-to-dev sync workflow. Covers merge strategy, CI architecture, failure investigation, and known issues.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[3262,3263],{"name":3075,"slug":3076,"type":15},{"name":3156,"slug":3157,"type":15},"2026-07-30T05:29:03.275638",496]