Start here: parse your first document in minutes.

Quickstart

1 · Create your account and organization

Sign up and create an organization. Everything in LangParse — documents, keys, billing — is scoped to an organization, and you choose its data-residency region (US or Australia).

2 · Get an API key

In the dashboard, open Settings → API keys and create one. New keys are scoped to parse by default (submit documents and read results); manage and admin scopes require the admin role. The secret is shown once — store it safely and use a separate key per environment so usage is attributable and revocation is surgical.

Send it on every request as a bearer token (or the X-Api-Key header):

Authorization: Bearer YOUR_API_KEY

3 · Find your model id

You parse against a model — a document type and the schema it returns. List the models in your organization to get an id:

curlcurl https://api.langparse.dev/api/models \
  -H "Authorization: Bearer $LANGPARSE_API_KEY"

4 · Parse a document

Send the file as base64 (≤ ~12 MB) to the model’s parse endpoint. For larger files, request a presigned upload first via POST /api/v1/models/{modelId}/uploads, PUT the bytes, then parse with the returned sourceKey.

curlcurl https://api.langparse.dev/api/v1/models/$MODEL_ID/parse \
  -H "Authorization: Bearer $LANGPARSE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "fileName": "invoice.pdf",
    "fileBase64": "JVBERi0xLjcK… (base64 of the file)"
  }'

Parsing is asynchronous — you get a documentId back right away:

202HTTP 202 Accepted
{
  "documentId": "doc_a1b2c3",
  "status": "queued",
  "pollUrl": "https://api.langparse.dev/api/v1/documents/doc_a1b2c3"
}

5 · Poll for the result

Fetch the document until status becomes processed (it moves through queued → processing → processed, or failed):

curlcurl https://api.langparse.dev/api/v1/documents/doc_a1b2c3 \
  -H "Authorization: Bearer $LANGPARSE_API_KEY"

The data object is your structured result; fields carries per-field confidence:

200{
  "documentId": "doc_a1b2c3",
  "modelId": "mdl_invoice",
  "status": "processed",
  "pageCount": 1,
  "data": {
    "vendor_name": "Acme Industries",
    "invoice_no": "INV-20418",
    "total_amount": 12480.00,
    "currency": "USD"
  },
  "fields": [
    { "name": "vendor_name",  "value": "Acme Industries", "confidence": "high" },
    { "name": "total_amount", "value": 12480.00,          "confidence": "high" }
  ]
}

Core concepts

Five ideas cover almost everything you’ll do with the API:

• Models — A model is a document type plus the schema you want back (the same thing the app UI calls a parser). You parse against a model id.
• Async by design — Parsing returns a documentId immediately (HTTP 202). Poll the document until status is processed — no long-held connections.
• Fields & confidence — Each extracted field comes back with a confidence of high, medium, low, or flagged, plus bounding boxes — so you can auto-accept or route to review.
• Routers — Classify an incoming file and split multi-document PDFs, then send each part to the right model automatically.
• Keys & scopes — API keys carry scopes — parse (submit + read results), manage (configure models/routers/sources), and admin. New keys are parse-only.

Where to go next

• Full API reference — every endpoint, parameter, and response.
• OpenAPI spec — generate a client/SDK or import into Postman.
• Enterprise — data residency, consensus, and BYO S3.
• Talk to us — stuck on something? We’re happy to help.