As February winds down, are your docs due for some spring cleaning?
We’ve just launched one of the most-requested improvements to our Self-Documenting Pull Requests: Now Dosu can maintain your existing documentation on 3rd-party platforms! You can import documents from GitHub, GitLab, Notion or Confluence. We’ll keep them up to date directly where they already live.
When managing docs on GitHub or GitLab, Dosu will create a PR (or MR) with its proposed changes; on other platforms, it will update the docs as soon as code changes are shipped.
Our goal is for this to work automatically in the background, but if you want to manually review and accept Dosu’s proposed changes, you can turn off the “Auto-Accept Review” option under the Space settings.
Make your coding agent faster, cheaper, and more aligned with your team in a single click
In case you missed our announcement earlier in the month, the Dosu MCP Server is now available to all users. TLDR: it improves speed, reduces cost, and increases the consistency of your coding agent’s output by keeping it aligned with your team’s standards and practices.
Ever notice your coding agent “researching the existing patterns in the codebase” over and over (and finding the wrong patterns anyway)? Our MCP Server doesn’t just give your agent access to the latest documentation, it also allows coding agents to automatically contribute documentation for the tasks they complete so they don’t have to start from scratch the next time you give them a similar task.
All you have to do to get started is head over to the app and follow the quick installation guide: one CLI command for Claude, one click for Cursor, and general instructions available for whatever else you're using.
If you want to see a demo of the Dosu MCP Server in action, check out our presentation from Auth0’s Camp AI event.
New Case Study: Bluefin
Three maintainers. 50,000 weekly users. 139 repositories. Bluefin's support queue was averaging a four-day response time, and the team had already split their Discord in half just to keep things readable. In our latest case study, Bluefin co-founder Jorge Castro walks through how they cut their community response times from 42 hours to 12 minutes, and why he hasn't written a systemd service unit in 6 months (spoiler: Dosu is handling the heavy lifting).
Bluefin is planning to ship Dosu's MCP server bundled with Goose directly in the OS, making it the first Linux distribution with built-in, documentation-aware AI assistance.
Tonight: Live in SF for a Fireside Chat
I’m excited to be speaking tonight at “From Bottleneck to Breakthrough: Why Context Engineering is the Missing Link in Software Velocity” from 6-8pm in San Francisco. There are still a few seats left!
I’ll be joined by leaders from Augment Code, LangChain and Innovation Endeavors to discuss when human engineering memory stops scaling, why agents fail without structured context, and what Dosu’s knowledge infrastructure can unlock for AI-powered teams. If you’re building with AI in production, this will be a candid, practical conversation, and a great room of engineering leaders shaping the next wave of AI-powered software.
March 26th: Live Panel with CodeRabbit, LanceDB and Continue
Join us next month in San Francisco for Infrastructure for Agentic Workloads: From Prototype to Production Without the Pain. We’ll have a panel of executive leaders building the next generation of AI-native devtools, from autonomous code review to living knowledge systems and local-first IDE agents.
Chang She — Co-Founder & CEO @ LanceDB
Devin Stein — CEO @ Dosu
Ty Dunn — Co-Founder & CPO @ Continue
Harjot Gill — CEO @ CodeRabbit
Seats are limited, so sign up today! We're looking forward to seeing you there.
Until next month,
Devin Stein
Founder and CEO