// Agent profile
Rowboat is an open-source AI coworker platform that orchestrates background agents across email, scheduling, and web tasks. It supports event-driven triggers (new email, daily schedule), connects to external tools via Model Context Protocol (MCP), and provides an isolated browser environment for safe web interactions. With 16K GitHub stars and active development through mid-2026, Rowboat occupies the 'AI coworker' niche — more autonomous than a chatbot, more structured than a general-purpose agent framework. The platform integrates with Claude Code, Codex, and other LLM backends, making it a flexible orchestration layer for teams building agentic workflows.
Rowboat is an open-source AI coworker platform built around multi-agent orchestration. Rather than requiring you to prompt an agent for every task, Rowboat lets you set up background agents that run autonomously — triggered by events like incoming email or on a schedule. Each agent has persistent memory across sessions and can connect to external services via MCP, giving it access to search, databases, CRMs, and custom internal tools.
Event-driven agents: Define triggers that launch agent workflows automatically. A new email arrives, an agent drafts a response using full work context. A daily schedule fires, an agent summarizes overnight activity. This moves agent interactions from pull (you ask) to push (agents act when conditions are met).
Browser isolation: Rowboat includes a sandboxed browser environment for web tasks. Because it runs separately from your main browser, you can grant it access only to specific accounts — keeping your primary browsing session private while the agent handles web-based workflows.
MCP integration: Any MCP-compatible server can be plugged into Rowboat as a tool. This means agents can interact with external APIs, search the web, query databases, or trigger automations without custom integration code.
Rowboat excels for teams that want autonomous agent workflows without building infrastructure from scratch. Email triage, daily reporting, web research, CRM updates, and cross-tool automation are natural fits. The event-driven model is particularly valuable for high-volume communication workflows where manual prompting would be impractical.
Rowboat is infrastructure-heavy compared to simpler agent tools — it requires self-hosting and configuration of triggers, MCP servers, and browser environments. Teams that just need a coding assistant or conversational AI should look at lighter tools. Rowboat is for teams ready to deploy autonomous agents as persistent background workers.
Engineering and operations teams building autonomous agentic workflows. Particularly valuable for teams already using MCP-compatible tools who want an orchestration layer that ties agents, tools, and triggers together into persistent automated workflows.
Opinionated bundle of agent personas — pre-built role templates for engineering, marketing, ops, and research teams that ship as drop-in markdown files.
AI-powered job search system built on Claude Code — 14 skill modes that evaluate listings against your CV, generate ATS-optimized resumes, and track your pipeline in a Go dashboard.
OpenAI's flagship AI assistant capable of conversation, analysis, coding, writing, and multimodal reasoning across text and images.