Docs for AI products, with updates drafted from what you merge.
Your models, endpoints and limits move fast. Workover drafts the docs update from what you merge, and serves every page as Markdown for the AI tools your users work in.
Yesterday’s quickstart, today’s broken build.
A new model or a changed rate limit can make yesterday’s quickstart wrong. Developers find out when their code breaks, and so do the coding assistants they ask.
Docs that move with the API.
Drafts from what you merge
When a pull request lands on GitHub, Workover finds the articles it affects and drafts the edits as suggestions.
AgentsAPI reference from your spec
OpenAPI, Swagger, AsyncAPI or GraphQL, read again on every publish when you give a spec URL.
API referenceMarkdown and llms.txt
Every page is also served as Markdown. llms.txt and llms-full.txt are generated on every publish.
Found by AI tools
Markdown on request, skill.md and .well-known discovery files, so AI tools find and read your docs the standard ways.
Open it in the tool they use.
Every page has a Copy page menu. Readers copy it as Markdown, or open it in ChatGPT, Claude or Perplexity, or connect it to Cursor or VS Code.
See how much of your traffic is AI.
Analytics separate people from AI agents, and list what readers searched for and did not find. No cookies.
See it live on a working docs site.
Workover's own docs, published with Workover at workover.io/docs. Open workover.io/docs
Pro is $149 a month, with 5 seats included.
Billed yearly, or $179 billed monthly. Pro adds the agents, Health and private sites to Essentials. The agents run on your own AI key, so model usage is billed by your provider. Extra seats are $20 and extra sites $49 a month.
Questions
Yes, with two-way Git sync to GitHub or GitLab, in Mintlify’s repo layout. Developer documentation
Tell the agent in a sentence, with screenshots if you have them. It updates every affected article, as suggestions to review.
No. Readers get Markdown, llms.txt and the Copy page menu today. Your writers can use the Workover CLI and MCP server. CLI and MCP
Start with the docs you already have.
Import them from their address, connect GitHub, and review the first suggested updates in your own workspace.