Contact
Start a conversation with the Overshow team
Product questions, beta access, or an enterprise enquiry. Pick the path that matches what you need.
Enquiry
Tell us your context and priorities
FAQ
Common questions
17 answers across capture, privacy, and pricing questions.
Screens (every monitor you record), meeting audio when a meeting is active or manually started, and local documents in folders you point it at. OCR reads text off every captured frame. Transcription runs in fixed 20-30 second windows with a short overlap tail.
In a local SQLCipher-encrypted SQLite database on your laptop, with FTS5 for full-text search and EmbeddingGemma 300M vectors for semantic search. No screen images or video files are persisted: only OCR text, audio transcripts, and metadata derived at capture time. Nothing captured is sent to a shared cloud database. The security page has the full control list.
By default, no captured content leaves. What does sit in the cloud: account and billing data, and device registration metadata. That is it. Optional MCP integrations and manual exports only move data when you explicitly enable them. See local vs cloud for the side-by-side.