Create Workflow

We recommend creating workflows in the anyformat platform where you can visually configure fields, test with sample documents, and iterate faster. Once your workflow is ready, copy its ID and use it with the API to run workflows programmatically.

POST /v2/workflows/ creates a workflow from a strongly-typed graph. Use it to:

Configure parse-node settings (engine, agentic mode, effort, prompt hints)
Build parse-only workflows (no extract node — markdown only)
Route documents through classifiers or splitters to multiple extract nodes
Run simple linear parse → extract workflows

Request Body

Field	Type	Required	Description
`name`	string	Yes	Workflow name
`description`	string	No	Optional description
`nodes`	Node[]	Yes	At least one node; exactly one must be `type="parse"`
`edges`	Edge[]	No	Directed edges; empty for parse-only workflows

Parse Node

The entry-point node. Exactly one per workflow.

Field	Type	Default	Description
`id`	string	—	Stable identifier (e.g. `"parse_1"`)
`type`	`"parse"`	—	Discriminator
`engine`	`"Fast" \| "Performant"`	`"Fast"`	Parser engine
`mode`	`"standard" \| "agentic"`	`"standard"`	Use `"agentic"` for per-block LLM routing through typed strategies (text/table/figure)
`effort`	`"low" \| "mid" \| "accurate"`	`"mid"` when `mode="agentic"`; must be omitted otherwise	Effort preset for agentic mode
`prompt_hint`	string	`null`	Domain hint shown to the parser (e.g. “medical lab report, preserve numerics exactly”)
`figure_enhancement_enabled`	boolean	`false`	Extract structured descriptions of charts and images (extra LLM cost)
`visual_grounding_enabled`	boolean	`true`	Per-block visual parsing with bounding boxes

Extract Node

Pulls structured fields from upstream parsed content.

Field	Type	Required	Description
`id`	string	Yes	Stable identifier
`type`	`"extract"`	Yes	Discriminator
`extraction_schema`	object	Yes	`{ "fields": [...] }` — at least one field
`use_images`	boolean	No (default `false`)	Pass rendered page images to the extraction model

Field shape — see Field Types.

Classify / Splitter Nodes

Branch routing nodes — see Field Types and the splitter cookbook recipes.

Edges

Field	Type	Required	Description
`source`	string	Yes	Source node id
`target`	string	Yes	Target node id
`branch`	string	Conditional	Required when leaving a `classify` (category id) or `splitter` (rule id) node; forbidden otherwise

Examples

Linear parse → extract

curl -X POST 'https://api.anyformat.ai/v2/workflows/' \
  -H 'Content-Type: application/json' \
  -H 'Authorization: Bearer YOUR_API_KEY' \
  -d '{
    "name": "Invoice extractor",
    "nodes": [
      {"id": "parse_1", "type": "parse"},
      {
        "id": "extract_1",
        "type": "extract",
        "extraction_schema": {
          "fields": [
            {"name": "invoice_number", "description": "Invoice ID", "data_type": "string"},
            {"name": "total", "description": "Grand total", "data_type": "float"}
          ]
        }
      }
    ],
    "edges": [{"source": "parse_1", "target": "extract_1"}]
  }'

Parse-only with agentic mode

A workflow with just one parse node and no edges. Produces markdown only — no extraction.

curl -X POST 'https://api.anyformat.ai/v2/workflows/' \
  -H 'Content-Type: application/json' \
  -H 'Authorization: Bearer YOUR_API_KEY' \
  -d '{
    "name": "Agentic parse-only",
    "nodes": [
      {
        "id": "parse_1",
        "type": "parse",
        "mode": "agentic",
        "effort": "mid"
      }
    ],
    "edges": []
  }'

Agentic mode routes each block through a typed strategy (text / table / figure) using Gemini for routing. effort picks a model mix — "low" (fast/cheap), "mid" (balanced, default), "accurate" (best quality, slowest). Use the Agentic Parse to Markdown cookbook for an end-to-end recipe.

Classify-then-extract (branched)

curl -X POST 'https://api.anyformat.ai/v2/workflows/' \
  -H 'Content-Type: application/json' \
  -H 'Authorization: Bearer YOUR_API_KEY' \
  -d '{
    "name": "Multi-doc",
    "nodes": [
      {"id": "parse_1", "type": "parse"},
      {
        "id": "classify_1",
        "type": "classify",
        "categories": [
          {"id": "INVOICE", "name": "Invoice", "description": "Vendor invoice."},
          {"id": "RECEIPT", "name": "Receipt", "description": "POS receipt."}
        ]
      },
      {
        "id": "extract_invoice",
        "type": "extract",
        "extraction_schema": {"fields": [{"name": "vendor", "description": "Vendor name.", "data_type": "string"}]}
      },
      {
        "id": "extract_receipt",
        "type": "extract",
        "extraction_schema": {"fields": [{"name": "merchant", "description": "Merchant name.", "data_type": "string"}]}
      }
    ],
    "edges": [
      {"source": "parse_1", "target": "classify_1"},
      {"source": "classify_1", "target": "extract_invoice", "branch": "INVOICE"},
      {"source": "classify_1", "target": "extract_receipt", "branch": "RECEIPT"}
    ]
  }'

Topology Rules

The endpoint enforces graph correctness — if any rule is violated, the request returns 400 and nothing is persisted.

Rule	Message
Exactly one parse node	`workflow must contain exactly one \`parse` node`
Unique node ids	`duplicate node ids: [...]`
Edges reference existing nodes	`edge source X not in nodes` / `edge target Y not in nodes`
Edge predecessor compatibility	`extract does not accept extract as predecessor`
Branch routing	`edge from classify X requires \`branch` to be set`
No fan-out from non-routers	`node Y (extract) has 3 outgoing edges; only classify and splitter may fan out`

Response

201 Created returns the workflow resource:

{
  "id": "0686bb97-8c30-70f0-8000-97669e000eb8",
  "name": "Invoice extractor",
  "description": "",
  "created_at": "2026-05-11T10:00:00Z",
  "updated_at": "2026-05-11T10:00:00Z"
}

Save the id — you’ll reuse it when uploading documents and fetching results.

Next Steps

Agentic Parse to Markdown

End-to-end recipe: create → upload → results, with the new agentic parse mode

Parse-Only Workflow

Convert documents to markdown without extraction

Field Types

All supported data_type values for extraction fields

Response Formats

The shape of parse, extractions, splits, and classifications on the results endpoint

Authorizations

Authorization

string

header

required

API key issued from app.anyformat.ai/settings. Send as Authorization: Bearer <key>.

Body

application/json

Request schema for creating a workflow.

name

string

required

Workflow name

Minimum string length: 1

Example:

"Invoice Processing"

fields

required

Field definitions. Each entry's shape is determined by its data_type.

Minimum array length: 1

Show child attributes

Example:

[
  {
    "data_type": "string",
    "name": "invoice_number"
  }
]

description

string | null

default:""

Workflow description

Example:

"Extracts invoice number, vendor, total, and line items."

Response

Successful Response

A workflow defines the extraction template — what fields to extract from documents, their types, and validation rules.

string

required

Unique identifier of the workflow (UUID).

Example:

"0686bb97-8c30-70f0-8000-97669e000eb8"

name

string

required

Human-readable name of the workflow.

Example:

"Invoice Processing"

description

string | null

Optional description of what this workflow extracts.

Example:

"A workflow for processing invoices and retrieving invoice details."

created_at

string<date-time> | null

Timestamp when the workflow was created (ISO 8601).

updated_at

string<date-time> | null

Timestamp when the workflow was last modified (ISO 8601).

fields

Fields · object[] | null

List of extraction field definitions configured for this workflow. null if not yet configured.

Overview

SDKs

Endpoints

Request Body

Parse Node

Extract Node

Classify / Splitter Nodes

Edges

Examples

Linear parse → extract

Parse-only with agentic mode

Classify-then-extract (branched)

Topology Rules

Response

Next Steps

Agentic Parse to Markdown

Parse-Only Workflow

Field Types

Response Formats

Authorizations

Body

Response

Overview

SDKs

Endpoints

Documentation Index

​Request Body

​Parse Node

​Extract Node

​Classify / Splitter Nodes

​Edges

​Examples

​Linear parse → extract

​Parse-only with agentic mode

​Classify-then-extract (branched)

​Topology Rules

​Response

​Next Steps

Agentic Parse to Markdown

Parse-Only Workflow

Field Types

Response Formats

Authorizations

Body

Response

Request Body

Parse Node

Extract Node

Classify / Splitter Nodes

Edges

Examples

Linear parse → extract

Parse-only with agentic mode

Classify-then-extract (branched)

Topology Rules

Response

Next Steps