Unstructured in. Structured data out.
Orves ORIS transforms complex documents into structured, source-linked data through one authenticated API call. Async by design: submit the document, poll the job, fetch the result. Standalone API, in private pilot; the first layer, not the final destination.
# submit a document (multipart) POST /v1/oris file=@contract_2024.pdf # 202 Accepted { "id": "prs_9f2c41…", "object": "oris", "status": "queued" } # poll until completed GET /v1/oriss/prs_9f2c41… { "status": "completed", "result": { "json_url": "…/result" } }
Parsing is where AI systems start lying.
Most parsers flatten documents into a string. Structure, geometry, tables and provenance are destroyed before any model sees the content — and no downstream step can recover what parsing lost.
Structure, not soup.
Structured output
- headings
- tables
- reading order
What the pilot returns today: the document's structure preserved as clean Markdown.
Async jobs
- job id
- polling
- idempotency
202 on submit, poll the job, fetch the result. A failed parse is never charged.
Source-linked by design
- provenance
- versioned
- roadmap
Output stays tied to the document that produced it. Typed elements, coordinates and certificates are the published direction — announced only as they ship.
From input to governed output.
One response, exactly as the API returns it.
Left: the page. Right: what you receive — the document's structure preserved as clean, structured Markdown.
# 200 OK { "markdown": "# Q2 Report\n\n## Revenue by segment\n\n| Segment | Q2 |\n| --- | --- |\n| Cloud | 104.2 |\n| Devices | 61.8 |\n\nRenewal is scheduled for…" }
Response from the live pilot contract. Typed structure atoms with geometry and coordinates are the published direction — announced when they ship, never before.
Measured before claimed.
What we measure: reading order, table structure, text accuracy, form fields and long-document fidelity — and charts, where parsing abstains rather than guess.
Public results are released only after pre-registered evaluation — until then, no number appears on this site. How we benchmark →
Every artifact is verifiable.
Certification runs through everything Orves produces. Verification is designed to be public, reproducible and independent of an Orves account.
Integrate in minutes.
Everything is delivered through the managed API — no additional infrastructure to deploy.