A CV parser API your code — or your AI client — can call

Aptora exposes CV parsing, validation and profile management programmatically: as an MCP server that plugs into Claude, Cursor and ChatGPT, and over an authenticated HTTP API. You send a document or a draft, you get back a validated profile in the published AptoraCV schema.

Real output, not a mockup
create_cv          title, data      -> id, validation
update_cv          id, data         -> validation
validate_cv        data             -> ok | failing fields
get_cv_schema                       -> AptoraCV JSON Schema
list_jobs / get_job / apply_to_job
search_candidates / shortlist_candidate / publish_job

One schema, published and versioned

Every integration reads and writes the same AptoraCV document. The schema is public — you can fetch it programmatically via get_cv_schema and validate any draft with validate_cv before saving. No guessing what a field is called.

MCP first: your users' AI assistants are the client

Aptora is a native MCP server. An assistant connected once can create profiles, tidy them, check them against the schema, browse job listings and apply — through the user's own Aptora account, with the user's own permissions. Sign-in is a standard OAuth flow.

Validation that says what is wrong

A draft that violates the schema comes back with the failing fields named, not with a generic 400. That makes the API usable from a language model as well as from code: the error message is the instruction for the fix.

The pipeline behind the endpoint

The same extraction pipeline that powers the product sits behind the API: document reading, field mapping into AptoraCV, and capability detection with normalized skills. You integrate one endpoint and inherit all of it.

Frequently asked questions

How do I authenticate against the Aptora API?

Via OAuth on the user's Aptora account. MCP clients like Claude, Cursor and ChatGPT run this flow with one click; the connection acts with exactly the permissions of the signed-in user.

Which operations does the MCP server expose?

Profile operations (create, read, update, validate, schema), job operations (list, read, apply) and, for team workspaces, talent search, shortlisting and job publishing. The setup guide lists every tool with its inputs and outputs.

Can I validate a CV document without creating anything?

Yes. validate_cv checks any draft against the AptoraCV schema and names the failing fields, without saving a profile.

Where do I find the setup instructions?

The MCP integration page has copy-paste configuration for Claude, Cursor and ChatGPT, and the docs cover the setup end to end.

Related