Dosu

It's Dosu Drop Day! 🎃 🔌 Live context from connected MCP servers (beta) 🔔 Approve knowledge updates in Slack 🔎 Deeper research, with the slowest answers about 30 seconds sooner 🏠 A limited self-hosted trial when security review blocks hosted Dosu go.dosu.dev/drop-oct26-b...

Should Humans Review An AI Agent's Work? Our CEO's, Devin Stein, take: people are already overwhelmed with review. Reviewing small pieces of conflicting information is work nobody has time for. Let the agent's do it.

dosu.devStop Re-explaining Your Company to AI AgentsPydantic's Laís Carvalho, Douwe Maan and Michael Pfaffenberger with Dosu's Devin Stein and Taylor Dolezal on shared knowledge for internal AI agents.

See the full discussion here: dosu.dev/webinars/sto...

In our Vercel AI SDK replays, agents learned the same build rule across 59 sessions and 40 tickets. One Codex run traced a missing import to a dependency's missing compiled output. What knowledge should you save for the next agent? dosu.dev/blog/coding-...

We gave Claude Code (Opus 5.5) and Codex (GPT 6.1 Sol) the same 49 Vercel/AI tickets, and we found some interesting gaps. Check out the full story (with graphs)! go.dosu.dev/agent-cost-g...

Does your coding agent choose a skill from the task...or only when you name the skill? Try a fresh session without the conversation where you wrote the instructions. Our procedural memory guide explains what to check. dosu.dev/blog/procedu...

After internal tests, we saw three embedding models rank undated Celery and DBOS facts in the same order for June and December questions. Your coding agent needs the history behind a fact to judge when to apply it. dosu.dev/blog/semanti...

How much better would a coding agent actually perform if it had the right memories from the start? VibeMemBench tries to answer that before testing existing memory systems. Devin explains why that made the benchmark the best memory one so far.

dosu.devRanking AI Systems: AI Memory BenchmarksTaylor Dolezal, Devin Stein and Michael Mangus rank seven agent memory benchmarks, from LoCoMo to VibeMemBench, by how much they tell coding agent teams.

Most AI memory benchmarks do not test for what real coding agents actually need to be good at: learning across long histories of work. But our team thinks VibeMemBench does and it is why it sits alone in S-Tier for Michael👇️

Your agent runs pytest again. In the last session, it learned that make test-int starts the database the tests need. That lesson may still be saved, but never retrieved. Before correcting your agent again, check what made it into context. go.dosu.dev/episodic-mem...

Knowing something and knowing how to do something are very different kinds of memory. EvoMemBench tries to measure both. Devin breaks down why that makes it interesting for real-world agents and where it still falls short for coding.

What if AI memory systems look bad because we’re measuring the wrong thing? Task completion alone doesn’t tell us much anymore. Did the agent learn from past work? Repeat old mistakes? Need the same human correction again? Michael thinks our benchmarks need to change.

It's Dosu Drop Day! 🎒 For our back-to-school edition, Dosu does the studying. When a coding session ends, Dosu writes down what your team would otherwise rediscover. 📝 Session-end study 🧾 Knowledge read cards 💳 Credits that go further go.dosu.dev/drop-sep26-bsky

The team keeps working after the agent’s session ends. @devstein.bsky.social on how merged PRs and resolved conversations keep agent memory current. Recording + transcript: dosu.dev/webinars/age...

22 Sept

Who reviews agent memory when people can’t keep up? Michael Mangus, Head of Product at Dosu, on trusting AI to maintain memory—and making mistakes easy to correct. Recording + transcript: dosu.dev/webinars/age...

21 Sept

A summary of a summary keeps getting further from the source. Michael Mangus, Head of Product at Dosu, explains how repeated summaries lose details in agent memory. Recording + transcript: dosu.dev/webinars/age...

go.dosu.devHow to design a CLI login flow for coding agentsDesign a CLI login for coding agents with a ticket flow that pauses, resumes in a new process, and returns the next command in JSON. Learn how Dosu built it.

18 Sept

Wrong memory can be worse than no memory. @devstein.bsky.social explains why stale agent memory can undo the value of remembering anything at all. dosu.dev/webinars/age...

Dosu reposted

Taylor Dolezal

@onlydole.dev

4 Sept

Ever wonder what happens when you sign in to Netflix on your TV? You get shown a code, approve it on your phone, and never have to type in a password. Coding agents don't get that luxury, though. Learn about how we solved that problem in Dosu's CLI, and where we found our inspiration!

go.dosu.devYour coding agent budget pays for context, not codePlanning and gathering context account for two thirds of agent costs.

26 Aug

Claude Code moved about 198 context tokens through its window for every output token in our PR replay study. Codex moved 134. This PR replay measures token flow and modeled API-rate spend. Check out the full breakdown in our latest blog go.dosu.dev/agent-budget...

19 Aug

It's Dosu Drop Day! 🗄️ Your Claude Code, Codex, and Cursor logs hold answers you keep spending tokens on. Now, Dosu stores them. 1. Backfill your Library from agent logs 2. Configure Dosu by asking Dosu 3. Analytics on login 4. See what knowledge your agent used go.dosu.dev/drop-aug26-b...

go.dosu.devIntroducing Decant: Insights for your Claude Code and Codex sessionsClaude Code and Codex log every session you run. Decant turns those logs into real numbers for what your agents did and how much it cost, without leaving your machine.

12 Aug

We launched a new OSS project today called Decant! Decant reads the logs Claude Code and Codex write and splits each run into categories of Context, Planning, Code, and Communicating. We'd love to hear what you think and what you learn from your sessions. go.dosu.dev/decant-aug26-bsky

go.dosu.devJuly Dosu Drop: Addition by SubtractionNo more waitlist, simpler Knowledge Cache tools, and new stats on Dosu's impact.

22 July

It's Dosu Drop Day! ☀️ This month's theme is addition by subtraction: 🚪 Waitlist? Gone. 🧠 Simpler Knowledge Cache tools (read / write / review) 📊 New cache stats in Analytics (our hit rate: 91%) 🔷 Azure DevOps support See the full Drop here: go.dosu.dev/drop-jul26-b...

go.dosu.devJune Drop: Introducing Libraries and AgentsWe overhauled how you configure Dosu, inspired by a short story from 1941.

24 June

Borges, in 1941, wrote a story about a library containing every possible book...infinite knowledge, yet useless without organization. That's our inspiration for June's Dosu Drop! Spaces are now Libraries, plus Agents in Slack, Teams & GitHub. ☔ go.dosu.dev/oes8aW7

go.dosu.devMay Drop: New usage analytics to see Dosu's impactPlus: bulk doc generation, support for more formats, and agent-driven setup

27 May

The May Dosu Drop is live! 🚀 We’re kicking off summer with an all-new Analytics tab to track where the agent is active and what knowledge your team references most. Plus, we just unlocked bulk document generation! Ask Dosu for a big batch of docs in a single prompt. go.dosu.dev/drop-may27-b...

dosu.devMay Drop: New usage analytics to see Dosu's impactPlus: bulk doc generation, support for more formats, and agent-driven setup

27 May

Also in this drop... * Full import support for AsciiDoc (.adoc) and .markdown file extensions * Interactive Slack setup. Mention @dosu.dev in a new channel or DM to instantly map data connections • Run "npx @dosu/cli setup --agent" to let Claude or Cursor handle MCP setup completely hands-free

Dosu reposted

Fabian Hiller

@fabianhiller.com

15 May

If you run an open-source project and aren't using @dosu.dev, you're missing out! Dosu's responses to new issues are really on point, and it's great to have them reference other issues and PRs, and brainstorm within an issue to make better decisions.

go.dosu.devHow Fresh Are Your Docs? Score Documentation Freshness in CILearn how to score documentation freshness in CI pipelines to keep your docs accurate and up to date. Automate quality checks and catch outdated content before it reaches users.

14 May

Most engineers stopped reading docs ages ago. But now, agents are reading them. What if you had a way to measure freshness on every PR? Is it as easy as implementing three signals, plus a Claude Code layer? Let's find out together!

go.dosu.devApril Drop: Knowledge from hundreds of open source projects, now available via MCPPlus templates, the return of Slack DMs, an improved CLI, and more!

22 Apr

The April Dosu Drop is out! ☔ Our MCP Server now includes knowledge from 800+ public OSS projects, like Airflow, Docling, vLLM, and more. Plus, we shipped templates, Slack DMs, and an upgraded CLI experience. Check out the full list here: go.dosu.dev/drop-apr26-b...

go.dosu.devA Claude Code Skill for Auto-Generating Project DocsGenerate your project's entire documentation layer from source code with a single Claude Code skill. Build /doc-it and run it against a real FastAPI project.

26 Mar

We built a Claude Code skill that generates project docs from source code. Then, we tested it on a FastAPI project. It caught stale AGENTS.md refs on the first pass and generated five files without inventing things it couldn't verify. Learn how to build your own: go.dosu.dev/doc-it-bsky

dosu.devNew at Dosu: Doc Export, Easier Reviews and Upcoming EventsGenerate docs with Dosu and sync them to Notion, Confluence or GitHub

18 Mar

It's Dosu Drop Day! Generate docs in Dosu, then you can export them to GitHub, Confluence, or Notion. Dosu keeps them maintained going forward. Last month's focus was import in, now this month is all about export out. dosu.dev/blog/march-2...

18 Mar

Also in this drop -> Doc Reviews get quite the upgrade. Edit proposed changes in place, see change summaries at a glance, and better quality updates across the board. Plus a new blog post on catching documentation drift with Claude Code and GitHub Actions.

luma.comAgentic Infrastructure: From Prototype to Production Without the Pain · LumaAgentic AI is pushing developer tools into a new era where systems don’t just assist, but reason, adapt, and act within real engineering workflows. Moving from…

18 Mar

If you're in SF on March 26, we're hosting a panel on infrastructure for agentic workloads with LanceDB, crewAI, and CodeRabbit at Hanwha AI Center. RSVP soon, as seats are limited! luma.com/ga3zi7sx

6 Mar

One unspoken agreement among every engineering team is that someone will update the docs... later. So, we built a GitHub Actions workflow that catches doc drift on merge, and wrote down every sharp edge we hit. dosu.dev/blog/how-to-...

dosu.devFebruary Dosu Drop: Maintain docs directly in Notion, Confluence and moreLet Dosu wrangle your docs with our new Import and Sync feature

25 Feb

It's Dosu Drop Day! Import docs from Notion, Confluence, GitHub, or GitLab, and Dosu keeps them updated as your code changes. No more docs that go stale too soon after you write them. dosu.dev/blog/februar...

25 Feb

Also in this drop. MCP Server is available to everyone, one click for Cursor, one command for Claude. Plus a new case study where Bluefin went from 42-hour support response to 12 minutes across 139 repos!

dosu.devJanuary Dosu Drop: New Year, New DocsBetter docs review and a first look at our MCP Server

22 Jan

It's Dosu Drop day! 🎉 We have an MCP Server preview (your coding agents can access Dosu's knowledge), an improved diff UI, and we're searching for design partners to test Notion/Confluence/GitHub sync. Let us know if you're interested in taking a look! dosu.dev/blog/january-2026-dosu-drop

17 Dec 2025

It's Dosu Drop Day! ❄️ This month, we have doc review diffs, image uploads, style guidelines, shareable links, and a CNCF case study to share! dosu.dev/blog/december-2025-dosu-drop