Rowboat: Your Local AI Coworker That Remembers Everything

Rowboat is an open-source AI coworker that transforms your work communications into a persistent knowledge graph, helping you get things done with context that actually remembers what matters.

What Rowboat Does

Most AI tools reconstruct context on demand by searching through documents or transcripts. Rowboat takes a fundamentally different approach - it maintains long-lived knowledge that accumulates over time. Your interactions with email, meeting notes, and other work data build a transparent "working memory" that compounds rather than starting cold every time.

The system connects to your existing work tools - Gmail for email, Granola and Fireflies for meeting notes - and automatically builds an Obsidian-compatible vault of plain Markdown notes with backlinks. This means you can inspect, edit, and update your knowledge graph anytime, seeing exactly how the AI connects different pieces of information.

Real-World Use Cases

Rowboat can handle tasks like:

• Meeting preparation: Pulls past decisions, open questions, and relevant threads into a crisp brief before your meeting with Alex
• Email drafting: Generates responses grounded in your history and commitments
• Document creation: Produces decks, docs, and PDF slides from your ongoing context
• Follow-ups: Captures decisions, action items, and owners so nothing gets dropped

Background Agents for Automation

Beyond on-demand assistance, Rowboat can spin up background agents to handle repeatable work automatically. You can configure it to:

• Draft email replies in the background using your past context
• Generate daily voice notes with agenda, priorities, and upcoming meetings
• Create recurring project updates from the latest communications
• Keep your knowledge graph current as new information arrives

You maintain complete control over what runs, when it runs, and what gets written back to your local Markdown vault.

Bring Your Own Model

Rowboat works with your preferred AI setup:

• Local models via Ollama or LM Studio
• Hosted models with your own API key/provider
• Model swapping anytime while your data stays in your local vault

Extensibility with MCP

The system supports Model Context Protocol (MCP), allowing you to connect external tools and services. This means you can integrate search engines, databases, CRMs, support tools, or your own internal systems. Supported integrations include Exa (web search), Twitter/X, ElevenLabs (voice), Slack, Linear/Jira, GitHub, and more.

Privacy-First Design

Everything lives on your machine as plain Markdown - no proprietary formats or hosted lock-in. You can inspect, edit, back up, or delete everything at any time. For voice notes, you can optionally add a Deepgram API key, but this remains entirely under your control.

Getting Started

Rowboat is available for Mac, Windows, and Linux. You can download the latest release from the GitHub repository or watch a demo video to see it in action.

For Google service integration (Gmail, Calendar, Drive), follow the Google setup instructions. Voice note functionality requires adding a Deepgram API key to ~/.rowboat/config/deepgram.json.

The project is actively developed with community support through Discord and Twitter. If you're looking for the Rowboat Web Studio, documentation is available in the repository.

Rowboat represents a shift from AI as a search tool to AI as a persistent coworker - one that remembers your context, understands what's relevant, and helps you act on it, all while keeping your data private and under your control.