Dovetail
AI-powered customer research repository for product teams
Local-first screen and audio recall that gives your AI agents context
screenpipe stands out for a genuinely local-first design: capture and storage default to the device, and the local API plus MCP server make the history usable by assistants like Claude and Codex instead of locking it in one app. That is a strong fit for privacy-conscious knowledge workers who want AI grounded in their real work. Two things to verify before committing. First, licensing: it is source-available under the Screenpipe Commercial License, so it is free for personal, non-commercial, educational, and research use, but commercial or production use requires a paid license, which is not the same as open source. Second, off-device data flow: optional cloud AI, transcription, sync, and integrations transmit selected context off the device, so confirm which features you enable. Continuous capture also raises real consent and storage considerations that teams should plan for. Verify current prices, included AI credits, and license terms with the vendor.
screenpipe is a local-first desktop app that captures searchable screen text and audio on your device and exposes it through a local API and MCP server so AI assistants can help with recall, meeting notes, summaries, and workflows. It is source-available under the Screenpipe Commercial License, with a free tier and paid individual, team, and enterprise plans.
screenpipe is a context layer for AI agents that runs on your own computer. It captures what appears on screen and what is heard and said, turning that stream into a searchable history of your work across apps, meetings, and conversations. Rather than recording every second indiscriminately, it captures meaningful moments and pairs each screen capture with the structured accessibility text the operating system already exposes, falling back to on-device OCR where that is unavailable. Speech is transcribed locally, with an option to use a cloud transcription provider instead. The product is built to feed AI tools rather than replace them. A local API and an MCP server expose the captured history so compatible assistants such as Claude and Codex can answer questions about your work, draft meeting notes, produce work summaries, and run scheduled workflows grounded in what screenpipe recorded. Recordings stay on the device by default, which keeps sensitive material local, while optional cloud AI, transcription, sync, and team features can transmit selected context off-device when a user chooses to enable them. The app runs on macOS, Windows, and Linux and is designed to work on typical machines from the last several years. screenpipe is source-available under the Screenpipe Commercial License rather than a traditional open-source license: personal, non-commercial, educational, and research use are free, and commercial or production use requires a paid license. The desktop product offers a free tier limited to one device with limited history, paid individual and team subscriptions that unlock full searchable history, cross-device sync, and team workspaces, and custom enterprise plans with controls such as single sign-on and device management. Buyers focused on privacy will value the local-first default, while those expecting a fully open-source license or fully free commercial use should read the license terms closely.
screenpipe is a local-first, source-available desktop app that captures searchable screen and audio history on your device and exposes it to AI agents through a local API and MCP server, with a free tier and paid individual, team, and enterprise plans.
screenpipe is developed by Negentropy Labs, Inc. and is a Y Combinator company from the S26 batch. The product is positioned as an open computer history and a local-first context layer for AI agents, letting users record their screen and audio continuously on their own device and provide that context to agents such as Claude and Codex.
The codebase is source-available and public on GitHub, and the company operates a commercial licensing model on top of it. Given the sensitivity of continuous capture, screenpipe emphasizes local-first storage and publishes security and trust materials that prospective buyers should review.
screenpipe captures screen content and audio, pairing screen captures with the operating system accessibility text and falling back to OCR where needed, and transcribing speech locally with an optional cloud provider. The result is a searchable history across apps, meetings, and conversations.
A local API and an MCP server expose that history to compatible AI assistants for recall, meeting notes, work summaries, and scheduled workflows. Recordings stay on the device by default, while optional cloud AI, transcription, sync, and integrations can share selected context off-device. It runs on macOS, Windows, and Linux, and higher tiers add cross-device sync, team workspaces with managed seats, capacity upgrades, and enterprise controls such as single sign-on and device management.
Privacy-conscious knowledge workers, consultants, researchers, and developers who want AI grounded in their real work, plus teams that need shared, managed access to captured work context.
Individuals who want searchable recall of their work and AI-generated notes and summaries.
Individual subscribers and team leads purchasing seats, plus enterprise buyers needing controls.
Privacy, local-first, and developer communities, and users of AI coding and agent tools.
A privacy-conscious professional or team that wants continuous, local-first work capture feeding AI agents through an open API and MCP, and is comfortable with a source-available commercial license.
screenpipe is built by Negentropy Labs, Inc. and is backed by Y Combinator as part of the S26 batch. Verify specific funding rounds and amounts via public sources.
It is a local-first desktop app that captures searchable screen text and audio on your device and exposes it to AI assistants through a local API and an MCP server.
Recordings are stored locally on your device by default. Optional cloud AI, transcription, sync, and integrations can transmit selected context off-device when enabled.
It is source-available under the Screenpipe Commercial License, not a traditional open-source license. Personal, non-commercial, educational, and research use are free, and commercial or production use requires a paid license.
Through its local API and MCP server it provides context to compatible assistants such as Claude and Codex, among other AI tools.
screenpipe runs on macOS, Windows, and Linux and is designed to work on typical machines from the last several years.
Side-by-side pages for pricing, features, and best-fit use cases.
AI-powered customer research repository for product teams
Energy-based AI daily planner that schedules tasks around your body's rhythms
Open-source local-first AI workforce coordinated by a Commander through one chat
Local-first memory for AI coding agents, via one MCP service