Process documents without operating document workers.
Upload a file, start an operation, and read back structured output. Flowyn runs the parsers and converters on managed workers and links every output to the job that produced it.
- job
- job_parse_7ab2
- attempt
- —
What teams build again and again
Every application that touches PDFs ends up running conversion binaries, parsers, upload handling, output storage, and cleanup somewhere other than the web request. That layer breaks in production long after the feature shipped.
What Flowyn provides
- 01PDF parsing into text, Markdown, or structural JSON
- 02DOCX parsing in document order
- 03HTML to PDF rendering in an isolated worker
- 04Intelligent splitting of mixed PDFs into child documents
- 05Outputs linked to the source file and job
What exactly happens
At the product-contract level. This is what your application and your agent can rely on.
- 01
Upload
Your request stores the operation and returns a job id with HTTP 202 before any processing starts.
- 02
Start
A managed worker claims the job, runs the operation, and records an attempt. Failures get up to 3 attempts.
- 03
Work
Outputs are stored as files attached to the job. Pages without a text layer stay empty; Flowyn does not perform OCR.
- 04
Read
You read the file job until it is complete, then download or read the output.
queuedprocessingretryingcompletefailedEverything around the process operation.
POST /v1/pdf/parsePDF Parse
Parse the text layer of a PDF into text, Markdown, or flowyn.pdf_parse.v1 JSON. Headings, tables, and reading order are preserved. Pages without a text layer stay empty.
ExplorePOST /v1/pdf/from-htmlHTML to PDF
An uploaded HTML file is rendered in an isolated browser worker and stored as a PDF linked to the job. Use it for invoices, statements, and reports generated from templates.
ExplorePOST /v1/docx/parseDOCX Parse
A DOCX is read in document order and returned as Markdown or flowyn.docx_parse.v1 JSON, keeping paragraphs, headings, lists, and tables in order.
ExplorePOST /v1/split-pdfIntelligent PDF Splitter
A mixed PDF is scanned for document boundaries, then split into child documents with page ranges and a confidence score for each boundary.
ExploreOne request. One resource you can read back.
Use an Idempotency-Key on durable create calls. Exact schemas are in the docs and OpenAPI.
For coding agents
A coding agent uploads the file, calls POST /v1/pdf/parse, keeps the job id, polls GET /v1/file-jobs/{id} until status is complete, then downloads the output file. Every field is described in the OpenAPI document.
curl -X POST https://api.flowyn.dev/v1/pdf/parse \
-H "Authorization: Bearer flw_test_••••" \
-H "Content-Type: application/json" \
-H "Idempotency-Key: parse-quarterly-001" \
-d '{
"file_id": "9c1e5b7a-2d4f-4a6b-8e3c-1f0a7d2b5c64",
"output": "markdown"
}'Also on Flowyn
Documents & files stays inside Flowyn's boundary: observe, verify, evaluate, execute bounded work, and report the result. The same account, API keys, and history cover every product.
Frequently asked questions
Does Flowyn perform OCR?
Is document processing synchronous?
Where do outputs live?
Start with documents & files.
Create a workspace, make your first call in the test environment, and inspect every attempt.