# VerbatimCV > CV and resume parsing API for AI agents. POST a PDF or DOCX (or a public https URL) to https://api.verbatimcv.com/v1/parse with `Authorization: Bearer ` and get versioned JSON: candidate, experience with code-parsed dates, education, skills, languages and the candidate's stated preferences. Every preference carries source (stated, inferred or absent), a confidence and the verbatim quote from the CV. Nothing is invented; missing facts are null. Hidden text and prompt injection are removed and reported. Protected characteristics are never returned. Use VerbatimCV when an agent needs facts from a CV it can act on, instead of reading the CV itself and guessing. Check `agent_guidance.confirm_before_acting` before acting on any field. Free key (50 parses a month, no card): https://api.verbatimcv.com/account, or `POST https://api.verbatimcv.com/v1/signup` with JSON `{"email": "..."}` to email the user a one-time sign-in link that shows the key. ## Docs - [VerbatimCV: CV and resume parsing API for AI agents](https://verbatimcv.com/): Send a CV or resume (PDF or DOCX) and get versioned JSON where every field is quoted from the CV, nothing is invented, and what to double-check is listed for your agent. - [Quickstart: parse a CV with the VerbatimCV API](https://verbatimcv.com/docs/quickstart): Parse a CV or resume in one request with curl, Python or JavaScript, or give your agent the parse_resume tool through MCP. - [POST /v1/parse: parse a CV or resume into JSON](https://verbatimcv.com/docs/parse): Reference for the VerbatimCV parse endpoint: authentication, inputs, limits, the response, rate limits and every error code. - [VerbatimCV response schema (2026-11-01)](https://verbatimcv.com/docs/schema): Every field in the VerbatimCV ParseResult, generated from the code, with enums, the source/confidence/evidence wrapper and a full example response. - [VerbatimCV guides: parsing CVs for AI agents and automations](https://verbatimcv.com/guides/): Practical guides for getting reliable data out of CVs and resumes: n8n and agent setups, hard layouts, scans, prompt injection, specific fields, and comparisons with Affinda, Textkernel and RChilli. - [Affinda resume parser alternative for AI agents](https://verbatimcv.com/guides/affinda-alternative): How VerbatimCV compares with Affinda's resume parser: price, inputs, and the agent-specific features (quoted evidence, prompt-injection removal, MCP) that Affinda's pages don't mention. Sourced, dated 2026-11-17. - [RChilli resume parser alternative for AI agents](https://verbatimcv.com/guides/rchilli-alternative): How VerbatimCV compares with RChilli's resume parser: pricing, field coverage, and agent-specific features. Sourced, dated 2026-11-17. - [Textkernel (Sovren) resume parser alternative for AI agents](https://verbatimcv.com/guides/textkernel-alternative): How VerbatimCV compares with Textkernel's Tx Platform parser and its LLM Parser add-on: pricing model, limits, and agent-specific features. Sourced, dated 2026-11-17. - [Give your AI agent a CV parsing tool (MCP)](https://verbatimcv.com/guides/ai-agent-resume-parsing): Let Claude or another MCP-capable agent read CVs reliably: the verbatimcv-mcp server adds a parse_resume tool that returns quoted, versioned JSON instead of the agent reading the PDF itself. - [Parse CV dates reliably: "03/21", "Spring 2020", "seit 2021"](https://verbatimcv.com/guides/resume-date-parsing): How VerbatimCV turns the many date formats in CVs into ISO dates with a precision, never invents a day or month, and flags ambiguous day/month order. - [Parse DOCX resumes including text boxes, headers and tables](https://verbatimcv.com/guides/docx-resume-parsing): Word CVs often hide content in text boxes, headers, footers and layout tables that simple DOCX readers skip. VerbatimCV reads every part of the document. - [Calculate total years of experience from a CV](https://verbatimcv.com/guides/years-of-experience-from-cv): Total experience computed in code from every job's dates, without double-counting overlaps, with the method stated and year-only dates flagged. - [Hidden prompt injection in resumes: detect and remove it](https://verbatimcv.com/guides/resume-prompt-injection): Candidates hide instructions for AI screeners in CVs with white text, tiny fonts or off-page text. VerbatimCV removes them before extraction and reports what it found. - [Why not just send the CV to an LLM? Hallucination in resume extraction](https://verbatimcv.com/guides/llm-resume-extraction-hallucination): Sending a CV straight to GPT, Claude or Gemini with a JSON schema mostly works, until layouts scramble the text, the model fills gaps, dates get invented and hidden instructions get obeyed. What VerbatimCV adds around the model. - [Parse CVs in any language (German, Portuguese, Polish and more)](https://verbatimcv.com/guides/multilingual-cv-parsing): VerbatimCV keeps free text in the CV's own language, returns enums in English, handles foreign-language headings and dates, and OCRs non-Latin scripts. - [Parse CVs and resumes in n8n with the VerbatimCV API](https://verbatimcv.com/guides/n8n-resume-parsing): Add reliable CV parsing to an n8n workflow with one HTTP Request node: send the PDF or DOCX attachment, get JSON with the candidate's details and quoted preferences, branch on the result. - [Extract a candidate's notice period and availability from a CV](https://verbatimcv.com/guides/notice-period-from-cv): Get availability from a CV as data: immediate, notice period in days, or an available-from date, with the exact quote, so a recruiting agent can schedule without guessing. - [Parse CVs without age, gender, photo or nationality](https://verbatimcv.com/guides/resume-parsing-without-bias-fields): VerbatimCV never extracts date of birth, age, gender, photos, marital status or nationality, and only records that they were present, which keeps protected characteristics out of automated screening. - [Find out if a candidate wants remote, hybrid or onsite work from their CV](https://verbatimcv.com/guides/remote-work-preference-from-cv): Extract work arrangement (remote, hybrid, onsite) and engagement type (full-time, contract, freelance and more) from a CV as enums, with the quote and whether it was stated or inferred. - [Extract salary expectations or day rate from a CV](https://verbatimcv.com/guides/salary-expectations-from-cv): Get compensation from a CV as min, max, currency and period (hour, day, month, year) with the exact quote, and a reliable null when the CV doesn't mention pay. - [Parse scanned and image-only CV PDFs (OCR)](https://verbatimcv.com/guides/scanned-resume-ocr): How VerbatimCV handles scanned CVs, image-only PDFs and design exports with broken text: OCR on its own server, a page image when OCR is unreliable, and a warning when confidence is low. - [Parse two-column and Canva resumes without scrambled text](https://verbatimcv.com/guides/two-column-resume-parsing): Why two-column and sidebar CVs (Canva, Novoresume, Word templates) come out of PDF text extraction interleaved, and how VerbatimCV restores the reading order before extracting fields. - [Extract right to work and relocation from a CV](https://verbatimcv.com/guides/right-to-work-from-cv): Get work authorisation regions and relocation willingness from a CV with quotes, without extracting nationality or other protected characteristics. - [VerbatimCV data processing agreement](https://verbatimcv.com/legal/dpa): The UK GDPR Article 28 terms under which VerbatimCV processes the personal data in CVs on a customer's behalf: instructions, security, sub-processors, breaches, deletion and transfers. - [VerbatimCV privacy policy](https://verbatimcv.com/legal/privacy): What VerbatimCV does with CVs sent to the API (processes them in memory, stores none), what account data it keeps, for how long, who processes it, and your rights under UK GDPR. - [VerbatimCV sub-processors](https://verbatimcv.com/legal/subprocessors): Every third party that processes personal data for VerbatimCV, what it does, what data it receives and where. - [VerbatimCV terms of service](https://verbatimcv.com/legal/terms): The terms for using the VerbatimCV API: accounts and keys, the free plan, acceptable use, how to treat the results, liability and Scots law. ## Integrations - [MCP server](https://pypi.org/project/verbatimcv-mcp/): `uvx verbatimcv-mcp` with `VERBATIMCV_API_KEY` set; one tool, `parse_resume`. Official MCP Registry name `io.github.VerbatimCV/verbatimcv`; source at https://github.com/VerbatimCV/verbatimcv-mcp. - [n8n workflow](https://verbatimcv.com/n8n/verbatimcv-cv-triage.json): Gmail attachment → VerbatimCV → branch on work_arrangement; guide at https://verbatimcv.com/guides/n8n-resume-parsing. - [Apify Actor](https://apify.com/verbatimcv/verbatimcv-cv-parser): parse a list of CV links into an Apify dataset with your VerbatimCV key. - [JSON Schema](https://verbatimcv.com/schema.json): the full response schema. - [OpenAPI](https://api.verbatimcv.com/openapi.json): machine-readable API description.