Dosu

Dosu is an automated knowledge capture platform and shared knowledge infrastructure designed for AI coding agents and development teams. It builds team-wide memory in the background during agent sessions, maintaining living documentation and syncing with various collaboration and code hosting tools. The platform helps development teams reduce token usage, keep implementations consistent between sessions, and improve agent performance without changing existing workflows.

• Automatic knowledge capture during coding agent sessions via the Dosu MCP Server
• Self-documenting pull requests that update documentation whenever code changes
• In-place synchronization with existing documentation platforms like Notion, Confluence, and Coda
• On-demand documentation generation through plain-text descriptions on Slack, MS Teams, and IDEs
• Support for custom document templates
• Observability and analytics dashboards for tracking team work and impact
• Team collaboration features allowing multiple users to share knowledge bases

Pros

• Reduces token costs and speeds up agent runs by providing consistent context to AI coding tools
• Automates documentation maintenance directly within existing code repositories and wikis
• Keeps humans in the loop with review and approval workflows for agent-generated activity
• Enables team knowledge sharing across multiple chat and productivity tools without requiring a migration

Cons

• The Free tier is limited to public repositories and provides 200 credits per month
• The Pro plan is restricted to 100 team members and 4000 credits per month
• Private repository connections and SaaS tool integrations require upgrading past the Free tier
• Advanced enterprise features such as SSO, custom RBAC policies, and audit logs are exclusive to the Enterprise tier

Dosu fits individual developers looking to optimize their coding agents as well as engineering teams seeking to capture internal knowledge, reduce bus factor risks, and improve AI agent accuracy. It is a poor fit for organizations that do not utilize AI coding agents or version control workflows. Additionally, teams requiring entirely on-premise deployments or custom integrations without an Enterprise tier arrangement should carefully review plan limitations.

Dosu uses a freemium pricing model with multiple tiers. The Free tier costs $0 per month, provides 200 credits per month, and is restricted to public repositories and a team member limit of 1. The Pro plan costs $16 per month, billed upfront at $192, and includes 4000 credits per month, private repository access, SaaS tool integrations, analytics dashboards, and a team member limit of 100. The Enterprise tier features custom pricing with credits as needed and unlimited team members, offering additional capabilities such as enhanced support, SSO, custom RBAC policies, and audit logs. A separate Maintainers Free tier is available at $0 for OSS maintainers, alongside a customized Maintainers Pro tier.

Unlike standard documentation tools that require manual authoring and maintenance, Dosu automatically captures knowledge in the background during AI coding agent sessions using an MCP Server. Rather than forcing teams to migrate their existing content, Dosu syncs and maintains documentation directly where it currently lives. It specifically targets the workflow of AI-assisted development, contrasting with traditional wikis by automatically updating files like AGENTS.md, CLAUDE.md, and skill definitions when code changes occur.

GitHub, GitLab, Slack, Confluence, Notion, Coda, MS Teams, Linear, Claude, Cursor, Codex, Gemini, and Windsurf

Dosu is worth trying for individual developers and engineering teams heavily reliant on AI coding agents who want to reduce token usage and automate documentation upkeep. The free tier allows users to test core features on public repositories without upfront cost, while the Pro plan unlocks private codebases and team collaboration features. Organizations requiring strict custom SLAs, dedicated support channels, and enterprise authentication will need to contact the sales team for a custom Enterprise plan. If a team does not use AI coding agents or collaborative chat platforms, the product offers limited utility.