About this project

Rowboat is a desktop AI coworker application available for Mac, Windows, and Linux. It indexes work data—including emails, meetings, Slack, and assistant conversations—into a living, Obsidian-style backlinked knowledge graph stored locally as plain Markdown. This context is used to power various built-in work surfaces. Key features include: - Brain: A knowledge graph that accumulates context over time, with explicit and inspectable relationships. - Email Client: Sorts emails by importance and automatically drafts responses using work context. - Background Agents: Agents that run on events or schedules, capable of connecting to tools, searching the web, using the browser, and writing code via Claude Code or Codex. - Built-in Browser: An isolated browser for collaborative web tasks with the AI. - Meeting Notes: A local note-taker that captures audio, produces live transcripts, and summarizes meetings into Markdown. - Code Mode: Spins up parallel coding agents driven by Rowboat's work context. - Apps: Custom work surfaces that can access tools, integrations, and be shared. - Integrations: One-click connections to popular products, plus support for external tools via Model Context Protocol (MCP). Rowboat supports bring-your-own-model setups, including local models via Ollama or LM Studio and hosted models via API keys. It is local-first by design, storing all data on the user's machine without proprietary formats or hosted lock-in. Optional integrations for voice input (Deepgram), voice output (ElevenLabs), and web search (Exa) are available via configuration files.