Hi HN,
I’m a co founder of Everdone (https://everdone.ai
).
We’re building a Work-as-a-Service platform — a place where teams trigger real work (not chats) and get production-ready outputs via AI. Everything is usage-based, like compute.
We’re starting with developer services because that’s where we’ve spent most of our lives as builders.
1. CodeDoc (live today)
AI documentation for GitHub repos — generated file-by-file, not guessed top-down.
How it works:
– Connect GitHub → choose repos → select files
– Instant documentation is generated
– Docs can be shared with unlimited team members
– You can set up automation at the repo/branch level
— runs daily
— only newly changed files are re-documented
— users can specify which files/ folders/ extensions to ignore
– We don’t train on customer code
Outputs explain what the file does, functions, DB operations, API I/O, dependencies, libraries used, etc. The tone is intentionally simple so junior developers can understand the project quickly.
We’ve documented a few OSS repos publicly (Next.js, FastAPI, Pandas, etc.) so you can see it in action.
Pricing:
– 50 files free
– $0.25 per file for AI documentation
– $0.025 per file/month for hosting and access
– Unlimited team members; no subscriptions or seat-based pricing
2. Coming next (developer side)
– CodeReview (AI-assisted code reviews)
– DBDoc (database documentation)
– ChangeLog (auto changelogs from Git diffs)
3. And beyond engineering
Long term, Everdone will support other departments:
content, analytics, reporting, QA, growth ops, sales ops, and more.
All usage-based — all producing real artifacts that organizations can audit and measure.
We think teams need a system of record for AI work. Not chat logs — actual outputs.
Would love feedback — technical, critical, or skeptical.
Happy to answer questions.
Comments URL: https://news.ycombinator.com/item?id=45952595
Points: 1
# Comments: 0
Source: everdone.ai
