Parse-only workflow

Nodes used: Parse A parse-only workflow is one Parse node and no edges. It returns the document as markdown plus one block per region, each with a position and a confidence, and runs no extraction. Use it to preview a document before you write fields, to feed your own LLM or search index, or to see how a document reads.

To parse one document on the Fast tier without creating a workflow, call POST /v3/parse/ (or POST /v3/parse/from-url/) with the file. It runs against a parse workflow that anyformat keeps for your organization and returns a run you poll like any other.

The graph

{
  "name": "Document parser",
  "nodes": [{ "id": "parse_1", "type": "parse", "mode": "standard" }],
  "edges": []
}

import os
from anyformat.sdk import Client

client = Client(api_key=os.environ["ANYFORMAT_API_KEY"])

workflow = client.workflow("Document parser").parse(mode="standard").create()

import { Anyformat } from "@anyformat/sdk";

const af = new Anyformat({ apiKey: process.env.ANYFORMAT_API_KEY! });

const workflow = await af.workflow("Document parser").parse({ mode: "standard" }).create();

mode selects the tier. Nothing else in the graph changes.

Pick a tier

Each row is what came back for the same one-page invoice.

Tier	`mode`	What runs	What you get back
Flash	`flash`	The PDF’s own text layer plus layout grounding. No model call. A page without a text layer is OCR’d (`scanned_pages`).	Plain text per block, tables as markdown pipe tables, `parse_confidence` of 100 on a born-digital PDF. Done in about 12 s.
Fast	`lite`	One OCR pass, no LLM correction.	One block per text line, tables as bare HTML `<table>`, `parse_confidence` and `layout_confidence` are `null`. Done in about 14 s.
Standard	`standard`	Page-by-page parsing with reading-order correction. The default.	Merged blocks with headings marked, HTML tables with a `data-cell-id` per cell, `parse_confidence` of 89. Done in about 28 s.
Agentic	`agentic`	Adaptive, multi-step parsing that works hardest on dense tables. `effort` sets the preset.	The Standard shape, with the most work spent on tables. Done in about 37 s. See Agentic parse to markdown.

Start on Standard. Move a document type down to Fast or Flash when its output holds up, and up to Agentic only where Standard falls short. The Parse node page lists the credits per page for each tier.

Run and read

# Upload one file and start a run
curl -X POST "https://api.anyformat.ai/v3/workflows/$WORKFLOW_ID/upload/run/" \
  -H "Authorization: Bearer $ANYFORMAT_API_KEY" \
  -F 'files=@document.pdf'
# -> 202 {"run_id": "...", "document_packet_id": "...", "status": "queued"}

# Read the run. Repeat until "status" is "processed".
curl "https://api.anyformat.ai/v3/runs/$RUN_ID/" \
  -H "Authorization: Bearer $ANYFORMAT_API_KEY"

result = workflow.run("document.pdf").wait()

print(result.parse.parse_confidence)
print(result.parse.markdown[:500])
for block in result.parse.blocks:
    print(block.id, block.type, block.page, block.parse_confidence)

import { readFile } from "node:fs/promises";

const file = new File([await readFile("document.pdf")], "document.pdf", { type: "application/pdf" });
const run = await workflow.run(file);
const result = await run.wait();

console.log(result.parse?.parseConfidence);
console.log(result.parse?.markdown?.slice(0, 500));
for (const block of result.parse?.blocks ?? []) {
  console.log(block.id, block.type, block.page, block.parse_confidence);
}

What comes back

The Standard tier on a one-page invoice, trimmed. extractions is empty because no Extract node ran.

{
  "id": "06a8d903-5ca9-79d9-8000-4dec88414c03",
  "status": "processed",
  "results": {
    "document_packet_id": "06a8d903-5a6d-796e-8000-770ffb2d85c5",
    "verification_url": "https://app.anyformat.ai/workflows/.../files/...",
    "parse": {
      "markdown": "<a id=\"p1_b0\"></a>\n\n# INVOICE\n\n<a id=\"p1_b1\"></a>\n\nInvoice #: INV-2026-9001 Issue date: 2026-02-20 ...\n\n<a id=\"p1_b5\"></a>\n\n<table>\n<thead>\n<tr>\n<th data-cell-id=\"r0c0\">Description</th> ...",
      "text": "INVOICE\nInvoice #: INV-2026-9001 Issue date: 2026-02-20 ...",
      "parse_confidence": 89.1,
      "layout_confidence": 0.7,
      "blocks": [
        {
          "id": "p1_b0",
          "type": "title",
          "page": 1,
          "bbox": { "x0": 0.115, "y0": 0.066, "x1": 0.269, "y1": 0.096 },
          "layout_confidence": 0.846,
          "parse_confidence": 100.0,
          "content": "# INVOICE",
          "hyperlinks": [],
          "rows": null,
          "image_base64": null
        },
        {
          "id": "p1_b5",
          "type": "table",
          "page": 1,
          "bbox": { "x0": 0.114, "y0": 0.317, "x1": 0.885, "y1": 0.409 },
          "layout_confidence": 0.697,
          "parse_confidence": 85.9,
          "content": "<table>...</table>",
          "hyperlinks": [],
          "rows": [
            [{ "cell_id": "r0c0", "text": "Description" }, { "cell_id": "r0c1", "text": "Qty" }],
            [{ "cell_id": "r1c0", "text": "Matrix Extensible Action-Items" }, { "cell_id": "r1c1", "text": "8" }]
          ]
        }
      ]
    },
    "classifications": [],
    "splits": [],
    "extractions": [],
    "edits": []
  }
}

Each <a id="p1_b0"> anchor in markdown names the block with the same id in blocks. That is how you map a passage back to its bbox (page fractions, 0 to 1) and its confidences: parse_confidence is 0 to 100, layout_confidence is 0 to 1.
A table block also carries rows, the cells as a grid, so you do not have to parse the HTML.
Block ids carry the page: p2_b0 is the first block on page 2. Every page is parsed.

Tips

Create one parse-only workflow and send every document to it. There is no reason to create one per document type.
Gate on parse_confidence. It is null on the Fast tier, which makes no model call, so fall back to a per-tier rule there.
For a folder of born-digital PDFs, Flash is the cheapest and fastest tier. A scanned page inside it is OCR’d by default; set scanned_pages to skip or fail if you would rather know.
The parse result is cached per file and tier. Send the same file again and the parse is reused; pass "cache": false on the node to force a fresh parse.

Next steps

Parse node

Every knob on the node, and the credits per tier

Agentic parse to markdown

The Agentic tier and its effort presets

Get run

The full run envelope

Quickstart

Add an Extract node and get fields back

Examples

Read only

Read, then pull out

Sort first, then pull out

Extract, then check

Operate at scale

Parse-only workflow

The graph

Pick a tier

Run and read

What comes back

Tips

Next steps

Parse node

Agentic parse to markdown

Get run

Quickstart

​The graph

​Pick a tier

​Run and read

​What comes back

​Tips

​Next steps

Parse node

Agentic parse to markdown

Get run

Quickstart

The graph

Pick a tier

Run and read

What comes back

Tips

Next steps