Feld API
The Feld API lets you programmatically upload documents, trigger classification, extraction and splitting, and retrieve structured results.
Base URL: https://<your-api-host>/v1
Current version: v1
The /v1 prefix is the same on every deployment — only the host changes. On the Feld cloud the host is api.feld.ai, so the base URL is https://api.feld.ai/v1, and that is the host used in the examples throughout these docs. If you were onboarded onto a dedicated or white-label deployment, your API host is provided together with your API key; use it with the same /v1 prefix.
Your API host is not your web app host
The address you use to log in to the web app is a different host, and it does not serve the API. Requests sent there never reach it — see wrong API host for how to recognise this.
What you can do
- Upload documents (PDF, Word, PowerPoint, Excel, images, text, emails)
- Classify documents against your configured schemas
- Extract structured data using your extraction schemas
- Split multi-page documents into sub-document boundaries
- Retrieve table data from documents
- Generate document summaries
- Manage reference data
Key concepts
| Concept | Description |
|---|---|
Authorization: Bearer | Header used for authentication. Each key is scoped to one project. |
document_id | UUID assigned to an uploaded document. Used in all subsequent operations. |
job_id | UUID tracking an async classification, extraction, splitting, or summary job. |
schema_id | Identifies which extraction or classification schema to apply. |
status (document) | Processing state of an uploaded document: pending → processing → completed or failed. |
status (job) | Processing state of an async job: pending → running → completed, failed or cancelled. |
metadata | Optional JSON key-value pairs attached to a document at upload time. |