VideoDB Skill

Perception + memory + actions for video, live streams, and desktop sessions.

When to use

Desktop Perception

• Start/stop a desktop session capturing screen, mic, and system audio
• Stream live context and store episodic session memory
• Run real-time alerts/triggers on what's spoken and what's happening on screen
• Produce session summaries, a searchable timeline, and playable evidence links

Video ingest + stream

• Ingest a file or URL and return a playable web stream link
• Transcode/normalize: codec, bitrate, fps, resolution, aspect ratio

Index + search (timestamps + evidence)

• Build visual, spoken, and keyword indexes
• Search and return exact moments with timestamps and playable evidence
• Auto-create clips from search results

Timeline editing + generation

• Subtitles: generate, translate, burn-in
• Overlays: text/image/branding, motion captions
• Audio: background music, voiceover, dubbing
• Programmatic composition and exports via timeline operations

Live streams (RTSP) + monitoring

• Connect RTSP/live feeds
• Run real-time visual and spoken understanding and emit events/alerts for monitoring workflows

How it works

Common inputs

• Local file path, public URL, or RTSP URL
• Desktop capture request: start / stop / summarize session
• Desired operations: get context for understanding, transcode spec, index spec, search query, clip ranges, timeline edits, alert rules

Common outputs

• Stream URL
• Search results with timestamps and evidence links
• Generated assets: subtitles, audio, images, clips
• Event/alert payloads for live streams
• Desktop session summaries and memory entries

Running Python code

Before running any VideoDB code, change to the project directory and load environment variables:

python
1from dotenv import load_dotenv
2load_dotenv(".env")
3
4import videodb
5conn = videodb.connect()

This reads VIDEO_DB_API_KEY from:

1. Environment (if already exported)
2. Project's .env file in current directory

If the key is missing, videodb.connect() raises AuthenticationError automatically.

Do NOT write a script file when a short inline command works.

When writing inline Python (python -c "..."), always use properly formatted code — use semicolons to separate statements and keep it readable. For anything longer than ~3 statements, use a heredoc instead:

bash
1python << 'EOF'
2from dotenv import load_dotenv
3load_dotenv(".env")
4
5import videodb
6conn = videodb.connect()
7coll = conn.get_collection()
8print(f"Videos: {len(coll.get_videos())}")
9EOF

Setup

When the user asks to "setup videodb" or similar:

1. Install SDK

bash
1pip install "videodb[capture]" python-dotenv

If videodb[capture] fails on Linux, install without the capture extra:

bash
1pip install videodb python-dotenv

2. Configure API key

The user must set VIDEO_DB_API_KEY using either method:

• Export in terminal (before starting Claude): export VIDEO_DB_API_KEY=your-key
• Project .env file: Save VIDEO_DB_API_KEY=your-key in the project's .env file

Get a free API key at console.videodb.io (50 free uploads, no credit card).

Do NOT read, write, or handle the API key yourself. Always let the user set it.

Quick Reference

Upload media

python
1# URL
2video = coll.upload(url="https://example.com/video.mp4")
3
4# YouTube
5video = coll.upload(url="https://www.youtube.com/watch?v=VIDEO_ID")
6
7# Local file
8video = coll.upload(file_path="/path/to/video.mp4")

Transcript + subtitle

python
1# force=True skips the error if the video is already indexed
2video.index_spoken_words(force=True)
3text = video.get_transcript_text()
4stream_url = video.add_subtitle()

Search inside videos

python
1from videodb.exceptions import InvalidRequestError
2
3video.index_spoken_words(force=True)
4
5# search() raises InvalidRequestError when no results are found.
6# Always wrap in try/except and treat "No results found" as empty.
7try:
8    results = video.search("product demo")
9    shots = results.get_shots()
10    stream_url = results.compile()
11except InvalidRequestError as e:
12    if "No results found" in str(e):
13        shots = []
14    else:
15        raise

Scene search

python
1import re
2from videodb import SearchType, IndexType, SceneExtractionType
3from videodb.exceptions import InvalidRequestError
4
5# index_scenes() has no force parameter — it raises an error if a scene
6# index already exists. Extract the existing index ID from the error.
7try:
8    scene_index_id = video.index_scenes(
9        extraction_type=SceneExtractionType.shot_based,
10        prompt="Describe the visual content in this scene.",
11    )
12except Exception as e:
13    match = re.search(r"id\s+([a-f0-9]+)", str(e))
14    if match:
15        scene_index_id = match.group(1)
16    else:
17        raise
18
19# Use score_threshold to filter low-relevance noise (recommended: 0.3+)
20try:
21    results = video.search(
22        query="person writing on a whiteboard",
23        search_type=SearchType.semantic,
24        index_type=IndexType.scene,
25        scene_index_id=scene_index_id,
26        score_threshold=0.3,
27    )
28    shots = results.get_shots()
29    stream_url = results.compile()
30except InvalidRequestError as e:
31    if "No results found" in str(e):
32        shots = []
33    else:
34        raise

Timeline editing

Important: Always validate timestamps before building a timeline:

• start must be >= 0 (negative values are silently accepted but produce broken output)
• start must be < end
• end must be <= video.length

python
1from videodb.timeline import Timeline
2from videodb.asset import VideoAsset, TextAsset, TextStyle
3
4timeline = Timeline(conn)
5timeline.add_inline(VideoAsset(asset_id=video.id, start=10, end=30))
6timeline.add_overlay(0, TextAsset(text="The End", duration=3, style=TextStyle(fontsize=36)))
7stream_url = timeline.generate_stream()

Transcode video (resolution / quality change)

python
1from videodb import TranscodeMode, VideoConfig, AudioConfig
2
3# Change resolution, quality, or aspect ratio server-side
4job_id = conn.transcode(
5    source="https://example.com/video.mp4",
6    callback_url="https://example.com/webhook",
7    mode=TranscodeMode.economy,
8    video_config=VideoConfig(resolution=720, quality=23, aspect_ratio="16:9"),
9    audio_config=AudioConfig(mute=False),
10)

Reframe aspect ratio (for social platforms)

Warning: reframe() is a slow server-side operation. For long videos it can take several minutes and may time out. Best practices:

• Always limit to a short segment using start/end when possible
• For full-length videos, use callback_url for async processing
• Trim the video on a Timeline first, then reframe the shorter result

python
1from videodb import ReframeMode
2
3# Always prefer reframing a short segment:
4reframed = video.reframe(start=0, end=60, target="vertical", mode=ReframeMode.smart)
5
6# Async reframe for full-length videos (returns None, result via webhook):
7video.reframe(target="vertical", callback_url="https://example.com/webhook")
8
9# Presets: "vertical" (9:16), "square" (1:1), "landscape" (16:9)
10reframed = video.reframe(start=0, end=60, target="square")
11
12# Custom dimensions
13reframed = video.reframe(start=0, end=60, target={"width": 1280, "height": 720})

Generative media

python
1image = coll.generate_image(
2    prompt="a sunset over mountains",
3    aspect_ratio="16:9",
4)

Error handling

python
1from videodb.exceptions import AuthenticationError, InvalidRequestError
2
3try:
4    conn = videodb.connect()
5except AuthenticationError:
6    print("Check your VIDEO_DB_API_KEY")
7
8try:
9    video = coll.upload(url="https://example.com/video.mp4")
10except InvalidRequestError as e:
11    print(f"Upload failed: {e}")

Common pitfalls

Examples

Canonical prompts

• "Start desktop capture and alert when a password field appears."
• "Record my session and produce an actionable summary when it ends."
• "Ingest this file and return a playable stream link."
• "Index this folder and find every scene with people, return timestamps."
• "Generate subtitles, burn them in, and add light background music."
• "Connect this RTSP URL and alert when a person enters the zone."

Screen Recording (Desktop Capture)

Use ws_listener.py to capture WebSocket events during recording sessions. Desktop capture supports macOS only.

Quick Start

1. Choose state dir: STATE_DIR="${VIDEODB_EVENTS_DIR:-$HOME/.local/state/videodb}"
2. Start listener: VIDEODB_EVENTS_DIR="$STATE_DIR" python scripts/ws_listener.py --clear "$STATE_DIR" &
3. Get WebSocket ID: cat "$STATE_DIR/videodb_ws_id"
4. Run capture code (see reference/capture.md for the full workflow)
5. Events written to: $STATE_DIR/videodb_events.jsonl

Use --clear whenever you start a fresh capture run so stale transcript and visual events do not leak into the new session.

Query Events

python
1import json
2import os
3import time
4from pathlib import Path
5
6events_dir = Path(os.environ.get("VIDEODB_EVENTS_DIR", Path.home() / ".local" / "state" / "videodb"))
7events_file = events_dir / "videodb_events.jsonl"
8events = []
9
10if events_file.exists():
11    with events_file.open(encoding="utf-8") as handle:
12        for line in handle:
13            try:
14                events.append(json.loads(line))
15            except json.JSONDecodeError:
16                continue
17
18transcripts = [e["data"]["text"] for e in events if e.get("channel") == "transcript"]
19cutoff = time.time() - 300
20recent_visual = [
21    e for e in events
22    if e.get("channel") == "visual_index" and e["unix_ts"] > cutoff
23]

Additional docs

Reference documentation is in the reference/ directory adjacent to this SKILL.md file. Use the Glob tool to locate it if needed.

• reference/api-reference.md - Complete VideoDB Python SDK API reference
• reference/search.md - In-depth guide to video search (spoken word and scene-based)
• reference/editor.md - Timeline editing, assets, and composition
• reference/streaming.md - HLS streaming and instant playback
• reference/generative.md - AI-powered media generation (images, video, audio)
• reference/rtstream.md - Live stream ingestion workflow (RTSP/RTMP)
• reference/rtstream-reference.md - RTStream SDK methods and AI pipelines
• reference/capture.md - Desktop capture workflow
• reference/capture-reference.md - Capture SDK and WebSocket events
• reference/use-cases.md - Common video processing patterns and examples

Do not use ffmpeg, moviepy, or local encoding tools when VideoDB supports the operation. The following are all handled server-side by VideoDB — trimming, combining clips, overlaying audio or music, adding subtitles, text/image overlays, transcoding, resolution changes, aspect-ratio conversion, resizing for platform requirements, transcription, and media generation. Only fall back to local tools for operations listed under Limitations in reference/editor.md (transitions, speed changes, crop/zoom, colour grading, volume mixing).

When to use what

Provenance

Reference material for this skill is vendored locally under skills/videodb/reference/. Use the local copies above instead of following external repository links at runtime.