domus-api

REST API that exposes the entire domus-captures documentation system as 43 programmable endpoints. 2,935 AsciiDoc files across 21 content categories, 1,107 partials, 495 examples, 21 standards, 53 projects, and 127 antora.yml attributes — all queryable as JSON. No database — the filesystem IS the database. FastAPI auto-generates OpenAPI schema, making the entire documentation system self-describing and machine-readable.

Every content category has a dedicated endpoint. Every skill (/incident, /project, /rca, /change-request) is a POST endpoint. Every document — pages, partials, examples — is queryable. The documentation system is an API.

A senior infrastructure engineer produces documentation every day. Worklogs, incident reports, change requests, root cause analyses, project plans, CLI patterns, security findings — thousands of documents across years of practice. The documentation lives in AsciiDoc files, rendered through Antora, deployed to a static site.

The site is readable. The content is structured. But the knowledge is trapped.

You cannot ask the system "what did I write about Mandiant?" without opening a browser and clicking through pages. You cannot create an incident report without switching to a terminal, remembering the right skill command, and hoping the template hasn’t changed. You cannot query your standards programmatically. You cannot let another tool — another machine, another agent — interact with your knowledge.

The documentation exists. The documentation is excellent. But it has no voice. It cannot speak to anything except a human with a browser.

domus-api exposes the documentation system as a REST API. Every document becomes a queryable resource. Every standard becomes a machine-readable contract. Every operational workflow — incident creation, change request scaffolding, root cause analysis — becomes a POST endpoint that produces files identical to what a human would create manually.

The filesystem is the database. There is no Postgres, no SQLite, no ORM. AsciiDoc files are the records. The directory structure is the schema. The Antora component is the data model. When you curl localhost:8080/search?q=mandiant, the API searches the same files you would search with grep. When you POST an incident, the API writes the same file your /incident skill would create.

Nothing is abstracted away. Nothing is duplicated. The API is a window into the filesystem, not a copy of it.