API reference

The identifier of the organization to charge for the parse. Optional when your account has exactly one Resume Parser organization, which is the usual case; required when it has more than one, so that the charge is never guessed. An identifier belonging to an organization you are not in, or to one of your organizations on the main Affinda platform rather than the Resume Parser, is answered 404 organization_not_found.

Only the literal multi selects the multilingual model. Every other value, including a specific language code such as fr, behaves as en. Omit it unless you want multi.

Names the file when the request body is the raw bytes of the document, as attachment; filename="cv.pdf". Ignored for multipart requests, where the file part carries its own name. The filename is a format hint and is not stored.

The parsed resume.

The request did not carry a usable file, or names no organization when it had to. missing_file means neither a file part nor a non-empty raw body was present, which includes a zero-byte upload and a body sent as application/x-www-form-urlencoded; organization_required means your account has several Resume Parser organizations and the organization query parameter has to say which one pays.

The API key is missing, malformed, unknown, expired, or blocked by your organization's IP allowlist. Also returned when a key from another region is used.

The organization has no parse credits left, or its credits have expired. The parser is not called and nothing is charged. Top up in the console's Billing screen.

Note this is a 402 on this API specifically. The same condition is a 403 on the main Affinda platform API.

The caller was recognised but not allowed. In practice this is an account that requires multi-factor authentication and has none set up; finish setting up MFA in the console.

The organization parameter names an organization that is not one of your Resume Parser organizations. An organization you are not a member of and one of your own platform organizations are answered identically, on purpose.

Only POST is implemented on this path.

The file is larger than the 20 MB limit. The size is checked against the declared Content-Length, against the multipart part, and again against the bytes actually received, so understating the length does not get a larger file through. The parser is not called and no credit is charged.

file_too_large is this API's own rejection. payload_too_large is the parser's, and reaches you only if a document slips past the first check.

The document is not a format the parser reads.

Intake is lenient: if a file might hold text, it is attempted. PDF, DOCX, ODT, XLSX, ODS, legacy .doc and .xls, RTF, HTML, SVG, plain text and the common image formats (PNG, JPEG, GIF, TIFF, BMP, WEBP) are all read. Legacy PowerPoint (.ppt) is the one format rejected outright. A self-hosted container with OCR switched off also answers this for images, because it has nothing to read them with.

The file was read but no resume could be extracted from it. A file with no text signal and a password-protected document both land here. Remove the password and retry in the second case. A zero-byte upload never reaches the parser and is answered 400 missing_file instead.

Too many requests. The limits are 6,000 parses per minute per organization and 6,000 per minute per API key, and the parser itself sheds load when its queue is full. Wait retryAfterSeconds and retry.

Something failed unexpectedly. The detail is recorded on our side, not returned.

The parser service could not be reached. The request is retried once on a fresh connection before this is returned, because a parse is idempotent and a dropped connection means it never arrived. No credit is charged.

The parser instance is shutting down. Retry.

The parse did not finish in time. The cloud waits 30 seconds for the parser, and the parser applies its own 28 second deadline, so in practice the parser's deadline is what you see. Unusually heavy documents may take longer than one attempt allows. No credit is charged.

The identifier of the organization to charge for the parse. Optional when your account has exactly one Resume Parser organization, which is the usual case; required when it has more than one, so that the charge is never guessed. An identifier belonging to an organization you are not in, or to one of your organizations on the main Affinda platform rather than the Resume Parser, is answered 404 organization_not_found.

Only the literal multi selects the multilingual model. Every other value, including a specific language code such as fr, behaves as en. Omit it unless you want multi.

Names the file when the request body is the raw bytes of the document, as attachment; filename="cv.pdf". Ignored for multipart requests, where the file part carries its own name. The filename is a format hint and is not stored.

The parsed resume.

The request did not carry a usable file, or names no organization when it had to. missing_file means neither a file part nor a non-empty raw body was present, which includes a zero-byte upload and a body sent as application/x-www-form-urlencoded; organization_required means your account has several Resume Parser organizations and the organization query parameter has to say which one pays.

The API key is missing, malformed, unknown, expired, or blocked by your organization's IP allowlist. Also returned when a key from another region is used.

The organization has no parse credits left, or its credits have expired. The parser is not called and nothing is charged. Top up in the console's Billing screen.

Note this is a 402 on this API specifically. The same condition is a 403 on the main Affinda platform API.

Name parts, date of birth, nationalities and location.

Emails, phone numbers and websites.

Totals derived from the work history, including tenure figures and employment gaps.

Base64 image, when one was found.