Your agent can read documents like databases.

01 THREE WAYS TO INTEGRATE

Pick the path that fits your stack.

curl -X POST https://api.talonic.com/v1/extract \
  -H "Authorization: Bearer $TALONIC_API_KEY" \
  -F "file=@invoice.pdf"

npm install @talonic/node

import { Talonic } from "@talonic/node";
const client = new Talonic({ apiKey: process.env.TALONIC_API_KEY! });
const result = await client.extract({ file_path: "invoice.pdf" });

npx @talonic/mcp

02 RESPONSE SHAPE

What your agent gets back

{
  "extraction_id": "ext_9f8a7b6c",
  "status": "complete",
  "data": {
    "rows": [
      {
        "vendor_name": {
          "value": "Acme Industrial GmbH",
          "confidence": 0.97,
          "phase": 1,
          "reasoning": "Exact match from field registry"
        },
        "contract_value": {
          "value": 248000.00,
          "confidence": 0.94,
          "phase": 2,
          "reasoning": "Extracted from clause 4.2"
        }
      }
    ]
  }
}

Every field carries confidence, phase, and reasoning. Your agent decides which values to trust.

03 COST TRANSPARENCY

Cost on every call

X-Talonic-Cost-Credits: 70
X-Talonic-Cost-EUR: 0.07
X-Talonic-Balance-Credits: 64930

GET /v1/credits/balance

{
  "balance_credits": 64930,
  "burn_rate_30d_credits": 12400,
  "projected_runway_days": 157
}

Track spend per call. Check runway before committing to batch jobs. Auto top-up is human-gated — your agent can request it but can't enable it.

04 FORMATS

Supported formats

25+ formats: PDF, DOCX, XLSX, PPTX, PNG, JPG, TXT, MD, HTML, JSON, CSV, XML, EML, ZIP, and more.

The API accepts multipart file upload or base64-encoded content. For batch processing, agents can submit up to 100 documents per job and poll for completion. Each format is handled by the same four-phase pipeline — PDF and scanned images go through OCR, while native text formats skip directly to field resolution.

06 SECURITY AND COMPLIANCE

Built for regulated industries

All document processing runs on EU-resident infrastructure (Azure Germany West Central). Documents are encrypted at rest and in transit, processed in ephemeral containers, and deleted after extraction completes. No document content is used for model training.

Talonic is certified under DIN SPEC 91491 for AI-based document extraction. The platform supports GDPR-compliant processing with data residency guarantees, audit logging, and role-based access control. HIPAA-eligible configurations are available for healthcare document workloads.