videodblisted

# VideoDB Python SDK Server-side video platform. Upload/ingest media, transcribe and index it, search for moments with timestamps, compose timelines, transcode and reframe, generate media with AI, ingest live streams, and record desktop sessions. Every operation returns HLS stream URLs compiled on demand — no local rendering. Trimming, concatenation, audio/music overlay, subtitles, text/image overlays, transcoding, resolution and aspect-ratio changes, transcription, and media generation are all handled server-side by VideoDB. Fall back to ffmpeg/local tools only for operations the timeline does not support: transitions, speed changes, crop/zoom, color grading, per-track volume mixing (see `references/editing.md`, Limitations). ## Setup ```bash pip install "videodb[capture]" python-dotenv # If the capture extra fails on Linux: pip install videodb python-dotenv ``` API key: `VIDEO_DB_API_KEY` as an environment variable or in the project `.env` file. Free key at console.videodb.io (50 free uploads). Let the user set the key; do not handle it directly. ```python from dotenv import load_dotenv load_dotenv(".env") import videodb conn = videodb.connect() # raises AuthenticationError when key is missing/invalid coll = conn.get_collection() # default collection ``` ## Core API map | Task | Call | |------|------| | Upload file/URL/YouTube | `coll.upload(url=... \| file_path=...)` -> `Video\|Audio\|Image` | | Transcript | `video.index_spoken_words(force=True)` then `video