Dosu for Individuals

Dosu auto-generates and maintains your documentation, specs, and agent memory — so you can execute on ambitious projects alone, keep your coding agents fast and cheap, and never lose context on work you did three months ago.

Free for 30 days. Just connect your GitHub repos and go

Build complex projects solo. Without losing your mind.

dosu — ~/backend

>

Sarah Chen

“Finally my READMEs are self-updating across all my repos. My agents are actually smart”

Sarah Chen

Individual builder

What Dosu does for individual builders

Automatic knowledge capture

Stop re-explaining your project. Every agent session means fewer tokens burned on rediscovery, fewer wrong turns, and consistent behavior run after run.

Documentation by default

Dosu gives you Self-Documenting PRs to maintain project docs automatically as the code evolves.

Create agent-friendly specs, skills, and memory.

When you spin up a Claude Code, Codex, or any other agent session, it already understands your project’s architecture, conventions, and current state.

Make your coding agent 2x faster at half the cost.

Coding agents slow down and burn tokens when they lack context. Dosu gives them exactly what they need.

Ask anything from anywhere

Documentation
ModelContextProtocolMCP Server
API References
Insights
SOURCES

# eng-backend

Sarah Chen11:55 AM

@Dosu how does our rate limiting work for /v2/ingest/endpoint?

DosuAPP11:55 AM

The /v2/ingest/endpoint uses a token bucket algorithm with per-org limits.

Rate Limiting - API Architecture

Confluence

Synced

Cloud Filestore here’s a quick breakdown:

📦 Instance2.5 TB BASIC_SSD
📁 Mount path/mnt/filestore
🔍 PowersLanceDB vector search
  • internal-api read-write (indexing)
  • search-api read-only (queries)
Ask Dosu anything…

dosu — ~/langchain

> what GCP service backs our code retrieval tools?

Found in infrastructure/gcp.tf and docs/architecture.md

Cloud Filestore (NFS) 2.5TB BASIC_SSD mounted at /mnt/filestore. Backs LanceDB vector DBs used by search-api (read-only) and internal-api (read-write).

Sources: [1] gcp.tf:42 [2] architecture.md:118

Set up in minutes, not days

1
1

Connect your GitHub repo

Authenticate with GitHub and point Dosu at your repository.

2
2

Dosu indexes your codebase

Dosu reads your code, issues, discussions, and existing docs to build a knowledge base of your project.

3
3

Your docs generate and stay current automatically

As your code evolves, Dosu keeps everything accurate — specs, memory files, READMEs, and more.

Built the way individual developers actually work.

You don’t have time to write documentation. You don’t have a team to maintain it. And you definitely don’t have budget for a tool that requires a sales call to get started. Dosu is free for public repositories that aren’t VC-backed or revenue-generating. Connect your GitHub repo in minutes and Dosu starts indexing immediately — no setup, no configuration, no documentation sprints required.

Frequently asked questions

Is Dosu really free?
Yes. Dosu is free for public GitHub repositories that are not VC-backed or primarily intended to generate revenue. If that's you, there's no catch.
What does Dosu actually generate?
Dosu generates and maintains READMEs, architecture docs, spec files, agent memory files, and skills documentation — all derived automatically from your codebase and kept current as it evolves.
Does it work with AI coding tools like Cursor and Copilot?
Yes. Dosu generates agent-friendly documentation and memory specifically designed to give tools like Cursor, Copilot, and other coding agents the context they need to perform better on your specific codebase.
What if I'm building something private?
Private repositories are also supported. Reach out or sign up to explore options for private or commercial projects.
How long does indexing take?
For most repositories, Dosu completes initial indexing within a few hours. Larger repos may take longer, but you'll be notified when it's ready.
Will Dosu answer questions about my dependencies too?
Yes. Dosu has knowledge of your open source dependencies, so it can answer questions that span your own code and the libraries it depends on.

You built something worth documenting

Let Dosu handle the part you’ve been putting off.