Barba-CV for AI Agents & Developers

Barba-CV is a deterministic CV transformation service exposed for programmatic use (including MCP-compatible automation flows).

It converts visually designed or inconsistent CV inputs into stable, machine-readable assets that can be reused across screening, tailoring, and recruitment workflows.

Core transformation model

  • Original structured CV: normalized JSON representation of extracted source content.
  • Generic ATS-compatible CV: flattened, recruiter-system-friendly version intended for broad compatibility.
  • Transformation diff: explicit comparison between source structure and ATS-ready output for traceability and downstream reasoning.

This triad is useful for automation because agents can inspect the before/after state, detect dropped or rewritten content, and apply policy checks with deterministic references.

Capabilities

  • CV parsing and normalization
  • ATS-oriented restructuring
  • Job-offer-aware tailoring flows
  • Deterministic JSON outputs for storage, QA, and replay
  • Export generation in multiple office formats

Output formats

  • JSON (canonical structured output from the API contract)
  • DOCX/PDF export artifacts when enabled in the publication workflow

How to think about integration

  • Treat Barba-CV as a normalization layer between unstructured candidate documents and downstream application logic.
  • Use the structured original CV as your canonical extracted dataset.
  • Use the ATS-compatible CV as your default submission-ready artifact.
  • Use the diff to drive auditability, approvals, or confidence scoring in agent pipelines.
  • Persist JSON identifiers in orchestration tools to support retries and deterministic post-processing.

Authentication and access

Requests require valid authorization according to the Barba-CV payment and access model. Implement explicit error handling and retry strategy in automation clients.

Endpoint overview

  • MCP gateway: POST /mcp (tools/list, tools/call)
  • Engine contract behind the gateway: POST /api/cv/process, GET /api/cv/status, GET /api/cv/results
  • Payment webhook endpoint: POST /stripe/webhook

Example MCP tool call

{  "method": "tools/call",  "params": {    "name": "barba.cv_process",    "arguments": {      "idempotency_key": "uuid-v4-key",      "cv_base64": "JVBERi0xLjc...",      "filename": "cv.pdf"    }  }}