Summaries
A summary is a short AI-written overview of a transcript, with the key points pulled out. It turns an hour of talking into something you can read in a minute — useful for deciding whether a video is worth watching in full, or for capturing the gist for your notes. You generate it from a transcript you already have.
A summary costs a flat 3 credits
Every summary costs 3 credits, whatever the length of the video — a five-minute clip and a two-hour talk cost the same. If generation fails, the credits are refunded automatically.
Summaries run on Gemini 2.5 Flash via the AssemblyAI EU LLM Gateway — an AI model reached through an EU-based gateway, so the transcript text stays in the EU while it's summarised.
Where it's saved, and editing it
The summary is saved with its transcript, so it's there whenever you open it in your library. You can edit the summary text, and — as with the transcript itself — your edits are kept separately from the AI's original version, which stays intact underneath.
Regenerating replaces the current summary
You can generate a fresh summary for the same transcript. Regenerating costs another 3 credits and overwrites the current summary — the previous one is not kept. INDXR asks you to confirm before it does.
Costs credits
Regenerating is a new 3-credit charge and replaces what's there. If you've edited a summary you want to keep, copy it out before regenerating.Sources
- INDXR (own code) — flat 3-credit cost regardless of length, refund on failureVerified against packages/shared/src/lib/pricing.ts (AI_SUMMARY); backend/main.py:1145-1156,1249,1179
- INDXR (own code) — model + EU gatewayVerified against packages/shared/src/lib/models.ts:41-43 (summaryModelName); backend/main.py:1126-1132
- INDXR (own code) — saved with transcript, edited kept separately, regenerate overwritesVerified against backend/main.py:1257-1269; apps/app/src/components/library/AiSummaryView.tsx:87-96; TranscriptViewer.tsx:1127
See also



