Agentic video primitives

Four layers behind one API, from the first frame to your agent. Use the ones you need, in our cloud or yours.

Add to Claude Get an API key

About 0.5 sfrom a question to the frames that answer it
100,000 hourschecked and cut into samples in under a week
300+ TBof video processed
1,000+developers and agents use VideoDB every week

Primitives

Files, buckets, cameras and screens go through one pipeline. Ask for any frame or window and get exactly that.

In: RTSP cameras, screens, microphones, files, links, S3, GCS and Azure buckets
Transcoding: In-house and distributed. Seeking costs the same anywhere in the file.
Frames: Any frame or window on request. Tiling packs a window into one image for a model.

See the architecture

Understanding

understand_video

{
  "segmentation": { "type": "time", "seconds": 2 },
  "analyzers": [{
    "type": "vlm",
    "name": "scene",
    "sampling": { "strategy": "uniform",
                  "frame_count": 3 },
    "config": { "model": "pro",
                "prompt": "Describe what is visible…" }
  }]
}

Done: 19 records, one for every 2 seconds of the homepage video

scene: 19 records, 6 shown

0:00–0:02
A small round robot in the foreground looks out over a misty mountain valley with a winding river and a cluster of illuminated village houses below…
0:06–0:08
A color landscape of snow-capped mountains and a misty valley, partially obscured by a semi-transparent, regular grid of colored dots…
0:12–0:14
Abstract, densely layered contour-like lines and swirls that look like a topographic map or fingerprint patterns…
0:22–0:24
Closely spaced curved contour lines that form a jagged, fractal edge on the left and broad sweeping arcs to the right…
0:28–0:30
A high-contrast black-and-white halftone pattern of evenly spaced dark dots on a pale background…
0:36–0:37.8
A small white spherical robot stands in the foreground overlooking a mist-filled mountain valley while the village windows glow…

On WGO-Bench, a VLM pipeline reaches 28.9% semantic F1 at IoU 0.5, against a 19.9% public baseline. Read the annotation benchmark

Indexes

4 moments from the scene index, best first

Moments that match the search, with their time in the video, an excerpt of the scene record and the score
0:00–0:02	…a winding river and a cluster of illuminated village houses below	0.45
0:02–0:04	…a winding river and a cluster of lit village houses, while jagged snow-capped peaks in the distance catch warm alpenglow	0.38
0:04–0:06	…warm pink/orange highlights on the summits and house windows	0.36
0:36–0:37.8	…bathed in warm alpenglow against a deep blue sky and the village windows glow	0.35

Answers come from five searches saved from real runs over the 37.8-second video at the top of the homepage, on Oct 30, 2026. Type your own words and the closest saved search answers.

VideoDB puts the right moment first 73.1% of the time, against 65.8% for a commercial video-native engine. Read the technical report

Delivery

python

q = "village houses with their lights on"
res = video.semantic_search(
    query=q, top_k=4,
    index_ids=[idx.index_id])

# every result is a playable stream
stream = res.compile()

MCPTools for any MCP client
SDKPython and Node
WebSocketEvents with the clip
WebhookWhen a job is done
ExportLeRobot, MCAP, Parquet

Run in our cloud or yours

Runs onAWSGoogle CloudAzureor your own cloud

SOC 2 Type II GDPR

Quickstart

Get your API key at console.videodb.io. Then upload a file, understand it, index it, search it and play the result.

pip install videodbPython

npm install videodbNode

npx skills add video-db/skillsCoding agents

Open the notebook Add to Claude

from videodb import connect

# uses your API key from console.videodb.io
conn = connect()
coll = conn.get_collection()
video = coll.upload(url="s3://lab/run-0412.mp4")

u = video.understand(analyzers=[{"type": "vlm", "name": "scene",
    "config": {"prompt": "What happens, and with what result?"}}])
u.wait_until_complete()
idx = video.index(name="scene", source=u.get_analyzer("scene"))

res = video.semantic_search(query="the gripper drops the bottle",
    index_ids=[idx.index_id], top_k=5)
for shot in res.get_shots():
    print(shot.start, shot.end, shot.metadata)

# the result, as a playable stream
stream = res.compile()