Upload anything
Drag in audio, video, images, markdown, or paste a YouTube URL. We support files up to 2 GB. Audio up to 500 MB, video up to 2 GB, images up to 25 MB.
The flow
From media to insight in five steps
Upload anything. We index it, chat with it, summarize it as structured insights, and let you share or monetize the result.
AI processes in seconds
Gemini transcribes audio and video. Gemma 4 OCRs images. Content is chunked into ~500-token segments and embedded into a 512-dim vector. Status updates live as your document moves through transcribing → chunking → embedding → ready.
Chat or search
Open the chat panel on any project. Ask anything. Get answers with citations that scrub-link to the exact moment in audio, video, or image. Top-K=8 vector retrieval is project-scoped — no cross-project leakage.
Generate insights
Click ✨ Insights on any project. Pick a starter (Glossary / Concept map / FAQ) or write your own prompt. The system retrieves the most relevant KB chunks and asks Gemini for structured output — markdown tables, recharts bar/pie, or D3 force-directed graphs. When KB coverage is thin, research mode transparently pulls authoritative public sources and flags them separately.
Share or monetize
Private, link-only, or public. Invite members with admin / editor / viewer roles. Insights inherit project visibility. Plan-tier monetization (paid creator KBs) is coming. For now, everything is free during the launch period.