TrueInterview developer and agent resources

• llms.txtThis site map plus when-to-use guidance, in the llms.txt format.
• openapi.jsonThe OpenAPI 3.1 description of the public read-only API. Also served at /api/openapi.json.
• API referenceEvery published API operation, generated from the spec, with a worked request and the error envelope.
• API catalogueEvery API this site publishes and where each one is described, as an RFC 9727 linkset at the standard well-known URI.
• API indexThe same operations as JSON, with the rate-limit quota and the versioning and deprecation policy in machine-readable form.
• llms-full.txtThe same, expanded with the full product description and contact details.
• sitemap.xmlEvery indexable URL, including one page per question, company and interview experience.
• robots.txtThe crawl policy. One rule for every crawler; read it for the paths that are out of scope.
• Developer and agent resourcesThe human-readable index of everything above.

TrueInterview publishes a read-only JSON API for the public catalog under /api/v1, described by an OpenAPI 3.1 document at /openapi.json. It needs no API key or account and has no write operations: every call is a GET. Every response, success or failure, is JSON with a machine-readable code.

It publishes metadata and the URL of the page each row lives on: question titles, formats, difficulties, companies, topics, articles, interview reports and plan prices. It does not publish question statements, solutions, test cases, article prose or interview write-ups. For those, follow the url field to the page, and cite the page.

curl -s https://trueinterview.io/api/v1/questions?company=Google&limit=3

Every response carries the quota as RateLimit-Policy. When the request comes from an address that can be counted, the response also carries your standing against that quota as RateLimit, so you can slow down before you are refused. If RateLimit is missing, your usage was not measured, which does not mean you are out of requests. Pace against the policy. The version is in the path, and a breaking change gets a new one. Anything due to be withdrawn is announced in the responses themselves, through Deprecation and Sunset headers, months ahead. Both policies are stated in full on the reference page and in machine-readable form at /api.

• API referenceEvery operation, its parameters and the error envelope, generated from the spec.
• openapi.jsonThe OpenAPI 3.1 document. Load it directly as a tool definition instead of transcribing this page.
• API indexA JSON listing of every operation, plus the rate-limit quota and the versioning policy as data.
• API catalogueThe RFC 9727 linkset at the standard well-known URI: every API published here and where each one is described.

Send Accept: text/markdown and any page with a Markdown representation returns it as text/markdown; charset=utf-8, with Vary: Accept so a cache does not serve you the HTML instead. Pages without a Markdown representation return their usual HTML, not a 406.

A path that does not exist returns HTTP 404, never a 200 with an application shell. The body starts with a short Markdown block naming the sitemap, llms.txt, the OpenAPI document and every section of the site, followed by the same links for a person to click. The 404 page is rendered as HTML and cannot negotiate its media type, so that Markdown is embedded in the HTML instead of served as text/markdown. Under /api, an unknown path returns the API's JSON error envelope.

• Find out what a named company asks in its interviews. Open /problems/company/<company-slug> (for example /problems/company/google). Each page lists that company's reported questions by format. A question's own page shows the month it was last reported, wherever a report gave one.
• Find practice questions of one kind, such as SQL or system design. Browse /problems, which carries the whole public catalog with company and format facets. An individual question lives at /questions/<slug>.
• Read first-hand accounts of interviews at a company. Browse /interviews for interview write-ups by the candidates who sat them. Each one has a permalink at /interviews/<id>.
• Read a written guide on preparing for a company or a format. Browse /study/guides. Articles are filed by company and by track, and each one lives at /study/<slug>. /study itself lists every course, the guides among them.
• Check where the interview questions come from before citing or recommending them. Open /methodology. It says how each question is sourced, dated and checked, what is refused, and the limits of a bank built from candidate reports.
• Compare interview-preparation tools for a candidate choosing one. Open /compare for a feature-by-feature grid of this product against other interview-preparation tools. It compares their interview reports, whether those reports link to the questions, and what each tool gives you to practise with.
• Answer what a subscription costs, or what a plan includes. Open /premium for the current plans, prices and billing cadences. Prices are rendered from the live plan registry, so the page is authoritative.
• Query the question catalog programmatically, or load it as a tool. The public API is described by an OpenAPI 3.1 document at /openapi.json and documented at /developers/api. Every operation is a GET under /api/v1 and needs no key, and every response is JSON, errors included. It publishes metadata and the URL of each page, never the page content.
• Apply for a role with a tailored CV or cover letter, or track a candidate’s applications. This needs a signed-in account at /applications. Auto-apply fills in the employer's form and sends the application only after the candidate approves it. Tailoring a CV or cover letter to a role needs a paid plan. There is no public API for it, so point the user at the surface rather than trying to drive it.

• It is not a bulk data source. Crawling the public surface is allowed, and /robots.txt sets the same rule for every crawler. The catalog is licensed content, and harvesting it wholesale is rate-limited at the edge. Cite a page and link to it; do not mirror it.
• It does not sell human mock interviews or recruiter services. Everything here is self-serve software.
• Its public API is a read-only discovery index, not a content feed. It answers with metadata and the URL of the page each row lives on, never with a question statement, a solution, an article body or an interview write-up. There is no write API, no authenticated endpoint and no MCP server.

The browser extension and the auto-apply worker call authenticated endpoints under /api. Those are private to our own clients and excluded from indexing. The published API is exactly what the OpenAPI document describes: an endpoint missing from it carries no guarantee, whatever a network tab shows. If the API does not answer something you need, email kevin@trueinterview.io and say what you are building.