TrueParser accepts document ingestion through a high-performance HTTP API. The ingestion process is designed to be robust, handling both direct multipart uploads and raw streams.
Submission Flow
The primary endpoint for document ingestion is POST /api/v1/documents/parse.
Ingestion Contract
- Authentication: Requires a valid JWT access token issued by the Dashboard.
- Payload: Accepts a multipart file upload.
- Identification: You can optionally supply a
documentType (e.g., PDF, DWG, SHP_ZIP). If omitted, TrueParser will attempt to detect the format automatically.
Explicit Routing Requirements
Some technical formats require explicit hints to ensure accurate parsing:
- SQL: Requires a
sqlDialect (e.g., PostgreSQL, Snowflake).
- CSV: Requires a
csvRoute to determine if it should be treated as a spatial GIS dataset or a tabular Office document.
- PDF: Can accept a
pdfMode for specialized extraction (e.g., SingleColumn, MultiColumn).
Document Identifiers
You can provide your own documentId at submission time to maintain consistency with your internal systems. If you don’t provide one, TrueParser will automatically generate a unique GUID for the job.
[!NOTE]
documentId uniqueness is enforced only within the active 3-hour retention window.
The ingestion API allows you to attach custom metadata JSON to any job. This metadata is carried through the pipeline and included in the final parsed output, providing traceability for your downstream applications.
TrueParser is optimized for high-volume ingestion. To achieve maximum throughput:
- Streaming Ingestion: Pass raw streams directly to the API where possible.
- Zero-Disk Buffering: The engine processes data from memory to storage, avoiding expensive disk I/O operations on the system.
- Rate Limiting: Ingestion is subject to your plan’s concurrency and byte limits. Check your Usage in the dashboard for current thresholds.
Last modified on April 1, 2026
