Rowboat is an open-source, local-first AI coworker designed to enhance productivity by turning work into a structured knowledge graph. Its primary function is to help users manage and leverage their accumulated knowledge to streamline tasks, improve decision-making, and automate routine processes. Unlike AI tools that rely on on-demand context retrieval, Rowboat builds and maintains a long-lived knowledge base, allowing context to accumulate and relationships to become explicit and easily accessible. This approach aims to provide a more efficient and personalized experience, making it a valuable tool for individuals and teams seeking to optimize their workflow.

The core features of Rowboat revolve around its ability to connect to existing workflows and build a comprehensive knowledge graph. It integrates with email services like Gmail and meeting note platforms like Granola and Fireflies, automatically extracting relevant information and incorporating it into the user's knowledge base. This knowledge base is stored locally as an Obsidian-compatible vault of plain Markdown notes, providing transparency and allowing users to inspect, edit, and manage their data directly. This local-first design ensures data privacy and control, eliminating vendor lock-in and allowing users to back up or delete their information at any time.

Rowboat's functionality extends beyond simple information storage. It empowers users to perform various actions based on their accumulated knowledge. For instance, users can generate documents and presentations, such as creating a PDF deck about a project roadmap or preparing for a meeting by summarizing past decisions and relevant threads. It also facilitates email drafting, follow-up management, and the creation of voice memos that automatically capture and update key takeaways in the knowledge graph. The system is designed to understand the user's context and provide relevant information when needed, such as during email replies or document creation.

A key differentiator of Rowboat is its use of background agents. These agents automate repetitive tasks, such as drafting email replies, generating daily voice notes, and creating recurring project updates. Users have control over these agents, defining their schedules and the information they process. This automation capability further enhances productivity by freeing up users from mundane tasks and allowing them to focus on more strategic activities.

Rowboat is designed to be flexible and adaptable. It supports various model setups, including local models via Ollama or LM Studio and hosted models through user-provided API keys. This allows users to choose the model that best suits their needs and preferences. Furthermore, Rowboat can connect to external tools and services via the Model Context Protocol (MCP), enabling integration with search engines, databases, CRMs, and other automation tools. This extensibility allows users to customize Rowboat to fit their specific workflows and integrate it with their existing toolsets.

The purpose of Rowboat is to empower users to work smarter, not harder. By building a comprehensive and accessible knowledge graph, Rowboat aims to reduce the time spent on repetitive tasks, improve decision-making, and enhance overall productivity. Its local-first design prioritizes user privacy and control, making it a secure and reliable tool for managing and leveraging personal and professional knowledge. The open-source nature of the project encourages community contributions and ensures that the tool remains adaptable and relevant to the evolving needs of its users. Rowboat strives to be more than just an AI tool; it aims to be a true AI coworker, assisting users in every aspect of their work.