Profile

Advanced options

Max hits per batchNo limit

Pause after this many matches; “Show more” adds another batch.

Fetch concurrencyParallel transcript downloads while searching. Applies on next search.Flush interval (ms)UI update cadence during search. Applies on next search.

Search Chat

Ask AI · Anilyzer

Ask a question about the transcripts

Bring your own API key. The chat searches the transcripts in your browser, sends the relevant excerpts to your chosen AI, and answers with citations. Nothing is hosted here — your key and the requests stay between your browser and the provider. See Use with AI for other ways to use this archive.

Set up your AI provider

ProviderModel

API key

Remember my key in this browser

Search mode

Auto uses your model’s native tool-calling when supported and falls back to a text protocol otherwise. Switch to Scripted if searches misbehave.

Max answer length (tokens)How long an answer can be before the model cuts it off. Gemini, OpenAI, and Claude all support ~8k; raise it for big reports.Sweep batch size (videos per batch)How many videos a whole-corpus sweep reads per batch. Lower it if a sweep keeps hitting your provider’s rate limit.Requests per minute

Paces requests to ≈1s apart so a free-tier key stays under its per-minute limit. Blank uses the per-model default (50). It auto-lowers itself after a rate-limit (429).

Debug

A redacted snapshot of this chat plus the most recent API calls (finish reasons + payloads) for troubleshooting. Your API key is never included.

Your key is stored only in this browser and is sent directly to Anthropic (Claude) — never to this site. Requests are billed to your own account. Get a key →

Saved chats

ReportReport mode

Ask a question about the transcripts. The assistant searches for what it needs — refining as it reads — and answers with citations.