Save a video worth revisiting.
Add a YouTube source, then fetch its captions when available. Videx does not invent a transcript for a video without one.
A research library for video
A long talk holds one useful answer. Save the videos you care about, fetch available captions, search for a passage and keep your notes beside the source.
Your saved sources and notes stay in your library. Run a private local workspace, or sign in on an account-enabled instance.
The answer is easier to trust when you can return to the speaker's words.
THE RESEARCH LOOP
Keep the path back to the video, so a promising passage stays verifiable.
Add a YouTube source, then fetch its captions when available. Videx does not invent a transcript for a video without one.
Find matching titles and transcript passages in your saved collection, then read timestamped captions.
Keep your notes beside the source and return to the original video at the moment you choose.
A USEFUL LIBRARY, NOT ANOTHER TAB
Search your saved titles and available captions without a model key. The local operator workspace also supports optional provider-powered search.
Read available captions in Videx and open the original video at a selected timestamp.
In the local workspace, choose a focus for suggested topics and example searches. This changes the interface, not the underlying indexed sources.
Save notes with each video in your library. The local operator workspace also offers video groups and YouTube account sync.
GET STARTED
Use an authorized Videx source checkout, Bun and Docker. The local dependency stack keeps PostgreSQL and Redis on loopback; the core indexing flow does not require an AI key.
cd app
docker compose -f compose.local.yml up -d --wait postgres redis
cd backend
if [ ! -e .env ] && [ ! -L .env ]; then cp .env.example .env; fi
bun install --frozen-lockfile
bun run db:generate
bun run db:migrate
bun run devcd app/frontend
bun install --frozen-lockfile
VIDEX_LOCAL_API_PROXY=1 bun run devCheck the API at http://127.0.0.1:3000/health/ready, then open the library at http://127.0.0.1:4321/.
The local workspace has no per-user account boundary. Keep its ports on your own machine. YouTube account sync and AI features need separate credentials when you choose to use them.