> ## Documentation Index
> Fetch the complete documentation index at: https://docs.anyformat.ai/llms.txt
> Use this file to discover all available pages before exploring further.
# Create Workflow
> Create a workflow from a typed graph of parse / classify / splitter / extract nodes in a single atomic transaction
We recommend creating workflows in the [anyformat platform](https://app.anyformat.ai) 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** (standard or agentic mode, prompt hints, figure enhancement)
* **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 |
| `mode` | `"standard" \| "agentic"` | `"standard"` | Use `"agentic"` for per-block LLM routing through typed strategies (text/table/figure) |
| `prompt_hint` | string | `null` | Domain hint shown to the parser (e.g. "medical lab report, preserve numerics exactly") |
| `figure_enhancement` | boolean | `false` | Extract structured descriptions of charts and images (extra LLM cost) |
### 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 |
| `lookup_file_uploads` | object\[] | No | Inline reference files for [smart lookup](#smart-lookup), each `{ "filename": string, "content": "" }` |
| `lookup_files` | string\[] | No | Reference files already stored as S3 URIs (alternative to `lookup_file_uploads`) |
| `lookup_suggestion` | string | No | Free-form hint telling the matcher how to join the document against the reference file |
Field shape — see [Field Types](/concepts/field-types). Any field may set `"source": "smart_lookup"` to make it a [smart-lookup field](#smart-lookup) (resolved from a reference file rather than read from the document), or `"source": "lookup_if_missing"` to extract it too and let the lookup only fill empty slots; the default is `"extraction"`.
### Classify Node
Routes the document to one of `categories[]` based on an LLM verdict. Outgoing edges must set `branch` to a category `id`.
| Field | Type | Required | Description |
| ------------- | ------------ | -------- | ---------------------------------------------- |
| `id` | string | Yes | Stable identifier |
| `type` | `"classify"` | Yes | Discriminator |
| `categories` | Category\[] | Yes | At least one category |
| `user_prompt` | string | No | Optional prompt prefix shown to the classifier |
Each `Category`:
| Field | Type | Required | Description |
| ------------- | ------ | -------- | --------------------------------------------------------------------- |
| `id` | string | Yes | Stable category id; matched against an outgoing edge's `branch` value |
| `name` | string | Yes | Display name shown to the LLM |
| `description` | string | Yes | Free-form description shown to the LLM |
### Splitter Node
Partitions a multi-document file into per-rule sub-documents. Each rule fires an outgoing edge whose `branch` matches the rule `id`; the downstream `extract` node runs once per resulting partition.
| Field | Type | Required | Description |
| ------- | ------------ | -------- | ----------------- |
| `id` | string | Yes | Stable identifier |
| `type` | `"splitter"` | Yes | Discriminator |
| `rules` | Rule\[] | Yes | At least one rule |
Each `Rule`:
| Field | Type | Required | Description |
| --------------- | ------ | ----------------- | ------------------------------------------------------------------------- |
| `id` | string | Yes | Stable rule id; matched against an outgoing edge's `branch` value |
| `name` | string | Yes | Display name |
| `description` | string | Yes | Free-form description shown to the splitter model |
| `partition_key` | string | No (default `""`) | Optional key used to label resulting partitions; empty means routing-only |
### Validate Node
Runs after an `extract` node and emits per-rule validation results. Rules are either AI-evaluated (`kind="ai"` with a natural-language `description`) or deterministic (`kind="deterministic"` with a structured `check` evaluated in pure Python, no model call).
| Field | Type | Required | Description |
| ------- | ------------ | -------- | ----------------- |
| `id` | string | Yes | Stable identifier |
| `type` | `"validate"` | Yes | Discriminator |
| `rules` | Rule\[] | Yes | At least one rule |
Each `Rule`:
| Field | Type | Required | Description |
| --------------- | ------------------------- | ---------------------- | ------------------------------------------------------------------------------------------------ |
| `id` | string | Yes | Stable rule id; round-trips through `ValidationResult.rule_id` |
| `kind` | `"ai" \| "deterministic"` | No (default `"ai"`) | Evaluation strategy |
| `name` | string | No | Optional display name for the rule card |
| `description` | string | Conditional | Natural-language prompt; required when `kind="ai"`, forbidden when `kind="deterministic"` |
| `check` | Check | Conditional | Structured deterministic check; required when `kind="deterministic"`, forbidden when `kind="ai"` |
| `severity` | `"error" \| "warning"` | No (default `"error"`) | Severity of a violation |
| `source_fields` | string\[] | No | `persistent_id`s of fields the rule references |
Deterministic `check` shapes (discriminated on `type`). Every shape below except `expression` names its field operands by `persistent_id`; `expression` addresses fields by extracted name instead.
| `type` | Required fields | Optional fields | Meaning |
| ------------ | ------------------------------------------ | --------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| `range` | `field` | `min`, `max` (at least one) | Numeric `field` is within `[min, max]` |
| `date` | `field` | `earliest`, `latest` (ISO `YYYY-MM-DD` or the literal `"today"`) | Date `field` is within the inclusive bounds |
| `arithmetic` | `operands[]` (at least one item), `equals` | `operator` (`"sum" \| "subtract" \| "product"`, default `"sum"`), `tolerance` (default `0.0`) | `operator(operands)` equals the value of the `equals` field, within `tolerance` |
| `comparison` | `left`, `op`, `right` | — | `left right` holds. `op`: `"==" \| "!=" \| ">" \| ">=" \| "<" \| "<="`. `right` is `{"source": "field", "field": ""}` or `{"source": "literal", "value": }` |
| `one_of` | `field`, `allowed[]` (at least one item) | `case_sensitive` (default `false`) | `field`'s value is one of `allowed` |
| `regex` | `field`, `pattern` | — | `field` matches `pattern`. Evaluated with google-re2 (linear time); stdlib `re` syntax that isn't re2-safe is rejected |
| `required` | `field` | — | `field` is present (non-null, non-empty) |
| `expression` | `expression` (1–1000 chars) | — | A [CEL](https://github.com/google/cel-spec) expression over the whole extraction, bound to one root variable `data`. Fields are addressed by extracted name (`data.total`, `data.lineas.map(l, l.importe)`). Helpers on top of CEL: `num(v)`, `sum(list)`, `abs(x)`. Anything the evaluator cannot answer with a boolean is `inconclusive` |
Example — arithmetic check (`subtotal + tax == total`, ±1¢):
```json theme={null}
{
"id": "totals_balance",
"kind": "deterministic",
"check": {
"type": "arithmetic",
"operator": "sum",
"operands": ["pid_subtotal", "pid_tax"],
"equals": "pid_total",
"tolerance": 0.01
},
"source_fields": ["pid_subtotal", "pid_tax", "pid_total"]
}
```
### 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` or `splitter` node; forbidden otherwise. **Must equal a category/rule `.id`** on the source node — not its `.name`. Using `.name` is rejected with `400`. |
## Examples
### Linear parse → extract
```bash curl theme={null}
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"}]
}'
```
```python Python theme={null}
import requests
response = requests.post(
"https://api.anyformat.ai/v2/workflows/",
headers={"Authorization": "Bearer YOUR_API_KEY"},
json={
"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"}],
},
)
workflow_id = response.json()["id"]
```
### Parse-only with agentic mode
A workflow with just one `parse` node and no edges. Produces markdown only — no extraction.
```bash curl theme={null}
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"
}
],
"edges": []
}'
```
```python Python theme={null}
import requests
response = requests.post(
"https://api.anyformat.ai/v2/workflows/",
headers={"Authorization": "Bearer YOUR_API_KEY"},
json={
"name": "Agentic parse-only",
"nodes": [
{
"id": "parse_1",
"type": "parse",
"mode": "agentic",
}
],
"edges": [],
},
)
workflow_id = response.json()["id"]
```
**Agentic mode** routes each block through a typed strategy (text / table / figure) using Gemini for routing. The mode is binary — set `mode: "agentic"` and you get the full agentic pipeline. See the [Agentic Parse to Markdown](/guides/recipes/agentic-parse-to-markdown) recipe for an end-to-end walkthrough.
### Classify-then-extract (branched)
Each outgoing edge from a `classify` (or `splitter`) sets `branch` to a category/rule **`.id`** on the source node. `.name` is the display label shown to the LLM and is **not** accepted as a branch value — using it returns `400`. In the example below, `cat_invoice` is the id used for routing; `Invoice` is the label the classifier sees.
```bash curl theme={null}
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": "cat_invoice", "name": "Invoice", "description": "Vendor invoice."},
{"id": "cat_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": "cat_invoice"},
{"source": "classify_1", "target": "extract_receipt", "branch": "cat_receipt"}
]
}'
```
### Smart lookup
Enrich extracted values by matching them against a **reference file** instead of reading them off the page. Flag the looked-up field with `"source": "smart_lookup"`, attach the reference file (inline as base64 via `lookup_file_uploads`, or by S3 URI via `lookup_files`), and optionally describe the join with `lookup_suggestion`. See [Smart Lookup](/concepts/smart-lookup) for how matching works and the Studio UI equivalent of these fields. Here the model reads `vendor_name` from the document, then resolves the canonical `vendor_id` from a vendor catalog:
```bash curl theme={null}
curl -X POST 'https://api.anyformat.ai/v2/workflows/' \
-H 'Content-Type: application/json' \
-H 'Authorization: Bearer YOUR_API_KEY' \
-d '{
"name": "Invoice + vendor lookup",
"nodes": [
{"id": "parse_1", "type": "parse"},
{
"id": "extract_1",
"type": "extract",
"lookup_file_uploads": [
{"filename": "vendor_catalog.csv", "content": "dmVuZG9yX25hbWUsdmVuZG9yX2lkCkFjbWUgQ29ycCxWLTAwMDEK"}
],
"lookup_suggestion": "Match the extracted vendor_name against the vendor_name column and return the matching vendor_id.",
"extraction_schema": {
"fields": [
{"name": "vendor_name", "description": "Vendor as printed on the document", "data_type": "string"},
{"name": "total", "description": "Grand total", "data_type": "float"},
{"name": "vendor_id", "description": "Canonical vendor code from the catalog, joined on vendor_name", "data_type": "string", "source": "smart_lookup"}
]
}
}
],
"edges": [{"source": "parse_1", "target": "extract_1"}]
}'
```
`lookup_file_uploads[].content` is the **base64-encoded** bytes of the reference file (CSV/text). The lookup field (`vendor_id`) is returned alongside the extraction fields in the results payload — there is no separate response section for it. An extract node that has a `source: smart_lookup`/`lookup_if_missing` field but no reference file is rejected with `400`.
## 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\` |
| Branch matches source `.id` | `edge from classify 'classify_1' sets branch='Invoice', but no category on 'classify_1' has that id (valid ids: ['cat_invoice']). \`branch\` must match a category \`.id\`, not its \`.name\`.\` |
| 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:
```json theme={null}
{
"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
End-to-end recipe: create → upload → results, with the new agentic parse mode
Convert documents to markdown without extraction
All supported `data_type` values for extraction fields
The shape of `parse`, `extractions`, `splits`, and `classifications` on the results endpoint
## OpenAPI
````yaml POST /v2/workflows/
openapi: 3.1.0
info:
title: anyformat API
description: >
Document extraction and workflow automation API.
AnyFormat lets you define extraction workflows that pull structured data
from any document — PDFs, images, scanned files, or plain text. Upload a
file, run it through a workflow, and get back structured fields with
confidence scores and source evidence.
## Quick start
See the [Quick start
guide](https://docs.anyformat.ai/api-reference/introduction) for a
walkthrough.
## Authentication
All endpoints (except `/health/`) require a Bearer token in the
`Authorization` header:
```
Authorization: Bearer
```
Get your API key from
[app.anyformat.ai/api-key](https://app.anyformat.ai/api-key).
## Versioning
Endpoints are versioned by path prefix (`/v2/`, `/v3/`). Every response
under
a versioned prefix includes `X-API-Version`.
version: 2.0.0
servers:
- url: https://api.anyformat.ai
description: API server
security: []
tags:
- name: workflows
description: >-
Workflows define extraction templates — what fields to extract from
documents. Create workflows, upload files, run extractions, and fetch
results.
- name: files
description: >-
File collections group uploaded documents and track their extraction
progress. Upload files, check status, and retrieve extraction results.
- name: suggestions
description: >-
AI-powered schema and workflow suggestion endpoints for the AnyFormat
webapp. Authenticated with a Cognito JWT rather than an API key.
- name: document-packets
description: >-
Document packets group uploaded documents for a workflow (v3). Address a
packet directly to (re-)run its workflow.
- name: runs
description: >-
Extraction runs — the v3 first-class result entity. Each run is a single
execution of a workflow against a document packet; fetch its status and
results flat by run id.
- name: health
description: Health check endpoints to verify API availability.
paths:
/v2/workflows/:
post:
tags:
- workflows
summary: Create workflow
description: |-
Create a workflow from a strongly-typed graph (atomic).
Provide an explicit list of typed `nodes` (parse / classify / splitter /
extract) and `edges` between them. The full workflow — fields, nodes,
routing — is created in a single transaction.
operationId: v2_create_workflow
requestBody:
required: true
content:
application/json:
schema:
$ref: '#/components/schemas/WorkflowCreateRequest'
responses:
'201':
description: Successful Response
content:
application/json:
schema:
$ref: '#/components/schemas/WorkflowResponse'
headers:
X-RateLimit-Limit:
description: Requests allowed per window (60s).
schema:
type: integer
X-RateLimit-Remaining:
description: Requests remaining in the current window.
schema:
type: integer
X-RateLimit-Reset:
description: Seconds until the current window resets.
schema:
type: integer
'401':
description: Authentication failed — missing, invalid, or expired API key.
content:
application/json:
schema:
$ref: '#/components/schemas/ErrorEnvelope'
examples:
AUTH_FAILED:
value:
error: Authentication failed.
detail: Authentication failed.
error_code: AUTH_FAILED
retryable: false
request_id: req_abc123
'422':
description: Validation Error
content:
application/json:
schema:
$ref: '#/components/schemas/HTTPValidationError'
'429':
description: >-
Rate limited — retry after the `Retry-After` header. Default
windows: 60 requests/minute for extraction submission, 600
requests/minute for general endpoints.
headers:
Retry-After:
description: Seconds until the next request is allowed.
schema:
type: integer
X-RateLimit-Limit:
description: Requests allowed per window (60s).
schema:
type: integer
X-RateLimit-Remaining:
description: Requests remaining in the current window.
schema:
type: integer
X-RateLimit-Reset:
description: Seconds until the current window resets.
schema:
type: integer
content:
application/json:
schema:
$ref: '#/components/schemas/ErrorEnvelope'
examples:
RATE_LIMITED:
value:
error: Rate limit exceeded — retry after backoff.
detail: Rate limit exceeded — retry after backoff.
error_code: RATE_LIMITED
retryable: true
request_id: req_abc123
security:
- ApiKeyAuth: []
components:
schemas:
WorkflowCreateRequest:
properties:
name:
type: string
minLength: 1
title: Name
examples:
- Invoice or receipt
description:
anyOf:
- type: string
- type: 'null'
title: Description
default: ''
nodes:
items:
oneOf:
- $ref: '#/components/schemas/PublicParseNode'
- $ref: '#/components/schemas/ClassifyNode'
- $ref: '#/components/schemas/SplitterNode'
- $ref: '#/components/schemas/PublicExtractNode'
- $ref: '#/components/schemas/PublicValidateNode'
- $ref: '#/components/schemas/PublicIfElseNode'
- $ref: '#/components/schemas/PublicSlackAlertNode'
discriminator:
propertyName: type
mapping:
classify:
$ref: '#/components/schemas/ClassifyNode'
extract:
$ref: '#/components/schemas/PublicExtractNode'
if_else:
$ref: '#/components/schemas/PublicIfElseNode'
parse:
$ref: '#/components/schemas/PublicParseNode'
slack_alert:
$ref: '#/components/schemas/PublicSlackAlertNode'
splitter:
$ref: '#/components/schemas/SplitterNode'
validate:
$ref: '#/components/schemas/PublicValidateNode'
type: array
minItems: 1
title: Nodes
edges:
items:
$ref: '#/components/schemas/Edge'
type: array
title: Edges
additionalProperties: false
type: object
required:
- name
- nodes
title: WorkflowCreateRequest
description: >-
Public-surface workflow create body — typed graph of parse / classify /
splitter / extract / validate nodes.
A ``validate`` node carries ``rules``; each rule is either AI-evaluated
(``kind="ai"`` + a natural-language ``description``) or deterministic
(``kind="deterministic"`` + a structured ``check``, evaluated in pure
Python with no model call).
Uses ``PublicNode`` (public request models with only the fields callers
may set); the domain ``AnyNode`` also carries ``filter`` (worker-only)
and executor-injected / staff-only fields, so the public bodies pin a
stricter node schema. Call :meth:`to_domain` before validating topology
or forwarding to the backend.
examples:
- description: Pull the invoice number and total from each document.
edges:
- source: parse
target: extract
name: Invoice totals
nodes:
- id: parse
mode: standard
type: parse
- extraction_schema:
fields:
- data_type: string
description: Invoice number as printed on the document.
name: invoice_number
- data_type: float
description: Invoice total in the document's currency.
name: total_amount
id: extract
mode: standard
type: extract
WorkflowResponse:
properties:
id:
type: string
title: Id
description: Unique identifier of the workflow (UUID).
examples:
- 0686bb97-8c30-70f0-8000-97669e000eb8
name:
type: string
title: Name
description: Human-readable name of the workflow.
examples:
- Invoice Processing
description:
anyOf:
- type: string
- type: 'null'
title: Description
description: Optional description of what this workflow extracts.
examples:
- A workflow for processing invoices and retrieving invoice details.
created_at:
anyOf:
- type: string
format: date-time
- type: 'null'
title: Created At
description: Timestamp when the workflow was created (ISO 8601).
updated_at:
anyOf:
- type: string
format: date-time
- type: 'null'
title: Updated At
description: Timestamp when the workflow was last modified (ISO 8601).
fields:
anyOf:
- items:
additionalProperties: true
type: object
type: array
- type: 'null'
title: Fields
description: >-
List of extraction field definitions configured for this workflow.
`null` if not yet configured.
type: object
required:
- id
- name
title: WorkflowResponse
description: >-
A workflow defines the extraction template — what fields to extract from
documents, their types, and validation rules.
ErrorEnvelope:
type: object
title: ErrorEnvelope
description: >-
Uniform error response. `retryable=true` means the same request may
succeed on retry (with backoff); `retryable=false` means the caller must
change the request before retrying.
required:
- error
- detail
- error_code
- retryable
- request_id
properties:
error:
type: string
description: Short human-readable summary.
detail:
description: >-
Machine-readable detail. String for most errors; list of field
errors for validation failures.
oneOf:
- type: string
- type: array
items:
type: object
error_code:
type: string
description: Stable identifier for programmatic handling.
examples:
- VALIDATION_ERROR
- AUTH_FAILED
- NOT_FOUND
- PRECONDITION_FAILED
- RATE_LIMITED
- INTERNAL_ERROR
- GATEWAY_TIMEOUT
retryable:
type: boolean
description: True when the caller should retry (with backoff for 429/412/5xx).
request_id:
type: string
description: Correlates this response with server logs.
HTTPValidationError:
properties:
detail:
items:
$ref: '#/components/schemas/ValidationError'
type: array
title: Detail
type: object
title: HTTPValidationError
PublicParseNode:
properties:
id:
type: string
minLength: 1
title: Id
description: Stable identifier for this node within the graph.
type:
type: string
const: parse
title: Type
mode:
type: string
enum:
- standard
- agentic
- lite
title: Mode
description: >-
Parse tier. 'standard' (default): page-by-page parsing with
reading-order correction — the best quality/speed balance for most
documents and tables. 'agentic': adaptive parsing that works hardest
on dense tables (50+ rows, many columns) — highest fidelity on
complex documents; extra cost per page. 'lite' (shown as 'Fast' in
the app): OCR-first tier, fastest and cheapest — accurate OCR by
default; set skip_routing_review=False to add a correction pass when
reading order, figures, or text formatting matter.
default: standard
prompt_hint:
anyOf:
- type: string
- type: 'null'
title: Prompt Hint
description: >-
Free-form hint that biases the parse output. In lite mode it guides
the correction pass, so it requires skip_routing_review=False.
figure_enhancement:
type: boolean
title: Figure Enhancement
description: >-
Standard mode: extract structured data from charts and images —
extra processing per figure. Ignored in agentic and in lite (use
`figures` instead).
default: false
cache:
type: boolean
title: Cache
description: >-
Reuse parsed output when the same file was already parsed with the
same mode and settings — a hit skips the whole parse. Disable to
force a fresh parse.
default: true
effort:
type: string
enum:
- low
- mid
- accurate
title: Effort
description: >-
Agentic mode: quality/cost preset. 'low' = several times cheaper but
more errors on dense or low-contrast tables, 'mid' (default) =
balanced, 'accurate' = highest fidelity. (Distinct from lite
`ocr_effort`.)
default: mid
ocr_effort:
type: string
enum:
- medium
- high
title: Ocr Effort
description: >-
Lite mode: 'high' (default) = the strongest OCR — best accuracy on
complex layouts, tables, and handwriting. 'medium' = faster and
cheaper, for clean digital documents. (Distinct from agentic
`effort`.)
default: high
skip_routing_review:
type: boolean
title: Skip Routing Review
description: >-
Lite mode: emit the OCR output as-is (the default — fastest and
cheapest). Set False to add a correction pass that fixes reading
order and OCR errors.
default: true
figures:
type: boolean
title: Figures
description: >-
Lite mode: the correction pass reads data points off charts/figures
and reconstructs them as HTML tables (needs
skip_routing_review=False).
default: false
text_formatting:
type: boolean
title: Text Formatting
description: >-
Lite mode: the correction pass restores inline styling
(bold/italic/superscript), code, and LaTeX math (needs
skip_routing_review=False).
default: false
routing_effort:
anyOf:
- type: string
enum:
- minimal
- low
- medium
- high
- type: 'null'
title: Routing Effort
description: >-
Lite mode: effort of the correction pass's reading-order analysis.
Lower it to cut latency on simple layouts.
additionalProperties: false
type: object
required:
- id
- type
title: PublicParseNode
ClassifyNode:
properties:
id:
type: string
minLength: 1
title: Id
description: Stable identifier for this node within the graph.
type:
type: string
const: classify
title: Type
user_prompt:
anyOf:
- type: string
- type: 'null'
title: User Prompt
description: >-
Optional extra instructions inserted between the classifier's system
prompt and the document text.
categories:
items:
$ref: '#/components/schemas/ClassifyCategory'
type: array
minItems: 1
title: Categories
additionalProperties: false
type: object
required:
- id
- type
- categories
title: ClassifyNode
SplitterNode:
properties:
id:
type: string
minLength: 1
title: Id
description: Stable identifier for this node within the graph.
type:
type: string
const: splitter
title: Type
rules:
items:
$ref: '#/components/schemas/SplitterRule'
type: array
minItems: 1
title: Rules
additionalProperties: false
type: object
required:
- id
- type
- rules
title: SplitterNode
PublicExtractNode:
properties:
id:
type: string
minLength: 1
title: Id
description: Stable identifier for this node within the graph.
type:
type: string
const: extract
title: Type
mode:
type: string
enum:
- standard
- agentic
- lite
title: Mode
description: >-
Extraction tier. 'standard' (default): single-pass extraction — the
best balance for simple documents and layouts. 'agentic': multi-step
extraction that maps every table to the schema and reasons across
sections — for dense or cross-page tables; extra cost per page.
'lite' (shown as 'Fast' in the app): standard extraction, faster and
cheaper; quality may vary on complex layouts.
default: standard
extraction_schema:
$ref: '#/components/schemas/ExtractionSchema'
description: Schema for the fields this node extracts. Required.
lookup_files:
items:
type: string
type: array
title: Lookup Files
description: Smart-lookup reference document URIs persisted on the extract node.
lookup_suggestion:
anyOf:
- type: string
- type: 'null'
title: Lookup Suggestion
description: Free-form hint shown to the smart-lookup matcher.
lookup_reasoning_effort:
anyOf:
- type: string
enum:
- minimal
- low
- medium
- high
- type: 'null'
title: Lookup Reasoning Effort
description: >-
Reasoning effort for the smart-lookup matcher (gpt-5 family):
minimal | low | medium | high. None = model default. Higher effort
improves match reliability on noisy join keys at higher
latency/cost.
lookup_file_uploads:
items:
$ref: '#/components/schemas/LookupFileUpload'
type: array
title: Lookup File Uploads
description: >-
Inline lookup-file content for the typed create call. Each entry is
uploaded and the resulting URI is appended to ``lookup_files``; this
field is create-input only.
use_images:
type: boolean
title: Use Images
description: >-
Standard/lite modes: send each PDF page's rendered image alongside
its parsed text so the model can read layout the text missed — adds
vision cost on every page.
default: false
additionalProperties: false
type: object
required:
- id
- type
- extraction_schema
title: PublicExtractNode
PublicValidateNode:
properties:
id:
type: string
minLength: 1
title: Id
description: Stable identifier for this node within the graph.
type:
type: string
const: validate
title: Type
rules:
items:
$ref: '#/components/schemas/ValidationRule'
type: array
minItems: 1
title: Rules
additionalProperties: false
type: object
required:
- id
- type
- rules
title: PublicValidateNode
PublicIfElseNode:
properties:
id:
type: string
minLength: 1
title: Id
description: Stable identifier for this node within the graph.
type:
type: string
const: if_else
title: Type
condition:
oneOf:
- $ref: '#/components/schemas/RangeCheck'
- $ref: '#/components/schemas/DateCheck'
- $ref: '#/components/schemas/ArithmeticCheck'
- $ref: '#/components/schemas/ComparisonCheck'
- $ref: '#/components/schemas/OneOfCheck'
- $ref: '#/components/schemas/RegexCheck'
- $ref: '#/components/schemas/RequiredCheck'
- $ref: '#/components/schemas/ConfidenceCheck'
- $ref: '#/components/schemas/ExpressionCheck'
- $ref: '#/components/schemas/ValidationCheck'
- $ref: '#/components/schemas/AllOfCheck'
- $ref: '#/components/schemas/AnyOfCheck'
title: Condition
discriminator:
propertyName: type
mapping:
all_of:
$ref: '#/components/schemas/AllOfCheck'
any_of:
$ref: '#/components/schemas/AnyOfCheck'
arithmetic:
$ref: '#/components/schemas/ArithmeticCheck'
comparison:
$ref: '#/components/schemas/ComparisonCheck'
confidence:
$ref: '#/components/schemas/ConfidenceCheck'
date:
$ref: '#/components/schemas/DateCheck'
expression:
$ref: '#/components/schemas/ExpressionCheck'
one_of:
$ref: '#/components/schemas/OneOfCheck'
range:
$ref: '#/components/schemas/RangeCheck'
regex:
$ref: '#/components/schemas/RegexCheck'
required:
$ref: '#/components/schemas/RequiredCheck'
validation:
$ref: '#/components/schemas/ValidationCheck'
additionalProperties: false
type: object
required:
- id
- type
- condition
title: PublicIfElseNode
PublicSlackAlertNode:
properties:
id:
type: string
minLength: 1
title: Id
description: Stable identifier for this node within the graph.
type:
type: string
const: slack_alert
title: Type
channel_id:
type: string
minLength: 1
title: Channel Id
channel_name:
type: string
minLength: 1
title: Channel Name
message_template:
type: string
minLength: 1
title: Message Template
severity:
type: string
enum:
- info
- warning
- critical
title: Severity
default: info
additionalProperties: false
type: object
required:
- id
- type
- channel_id
- channel_name
- message_template
title: PublicSlackAlertNode
description: >-
Terminal alert-sink. ``field_index`` is executor-injected — omitted
here.
Edge:
properties:
source:
type: string
minLength: 1
title: Source
target:
type: string
minLength: 1
title: Target
branch:
anyOf:
- type: string
- type: 'null'
title: Branch
description: >-
Source-port label for branch routing. Required when leaving a
classify or splitter node by category/rule.
additionalProperties: false
type: object
required:
- source
- target
title: Edge
description: >-
A directed edge between two nodes. ``branch`` carries the source-port
label
used for routing out of ``classify`` (category id) or ``splitter`` (rule
id) nodes.
ValidationError:
properties:
loc:
items:
anyOf:
- type: string
- type: integer
type: array
title: Location
msg:
type: string
title: Message
type:
type: string
title: Error Type
input:
title: Input
ctx:
type: object
title: Context
type: object
required:
- loc
- msg
- type
title: ValidationError
ClassifyCategory:
properties:
id:
type: string
minLength: 1
title: Id
description: Stable category id used as the edge `branch` value when routing.
name:
type: string
minLength: 1
title: Name
description: >-
Display name shown to the LLM; classification and routing key off
it, so it must be unique across categories.
description:
type: string
minLength: 1
title: Description
description: Free-form description shown to the LLM.
additionalProperties: false
type: object
required:
- id
- name
- description
title: ClassifyCategory
SplitterRule:
properties:
id:
type: string
minLength: 1
title: Id
name:
type: string
minLength: 1
title: Name
description:
type: string
minLength: 1
title: Description
partition_key:
type: string
title: Partition Key
description: >-
Field that further partitions this category into separate
sub-documents (e.g. invoice_number): each distinct value becomes its
own split, extracted independently. Empty = routing-only, the whole
category flows on as one document.
default: ''
additionalProperties: false
type: object
required:
- id
- name
- description
title: SplitterRule
ExtractionSchema:
properties:
fields:
items:
oneOf:
- $ref: '#/components/schemas/StringField'
- $ref: '#/components/schemas/IntegerField'
- $ref: '#/components/schemas/FloatField'
- $ref: '#/components/schemas/BooleanField'
- $ref: '#/components/schemas/DateField'
- $ref: '#/components/schemas/DatetimeField'
- $ref: '#/components/schemas/EnumField'
- $ref: '#/components/schemas/MultiSelectField'
- $ref: '#/components/schemas/ObjectField'
discriminator:
propertyName: data_type
mapping:
boolean:
$ref: '#/components/schemas/BooleanField'
date:
$ref: '#/components/schemas/DateField'
datetime:
$ref: '#/components/schemas/DatetimeField'
enum:
$ref: '#/components/schemas/EnumField'
float:
$ref: '#/components/schemas/FloatField'
integer:
$ref: '#/components/schemas/IntegerField'
multi_select:
$ref: '#/components/schemas/MultiSelectField'
object:
$ref: '#/components/schemas/ObjectField'
string:
$ref: '#/components/schemas/StringField'
type: array
minItems: 1
title: Fields
description: Field definitions making up this extract's output.
additionalProperties: false
type: object
required:
- fields
title: ExtractionSchema
description: |-
Schema describing what an extract node produces.
Modeled as a typed object (not a bare ``list[AnyField]``) so metadata
can be added later (versioning, derived flags, output-shape hints)
without another wire-format break. At least one field is required.
LookupFileUpload:
properties:
filename:
type: string
minLength: 1
title: Filename
content:
type: string
minLength: 1
title: Content
description: Base64-encoded file bytes.
additionalProperties: false
type: object
required:
- filename
- content
title: LookupFileUpload
description: |-
Inline lookup-file content carried on the typed create call.
The backend reads ``filename`` + ``content`` (base64-encoded bytes),
uploads the file, and stores the resulting URI in
``ExtractNode.lookup_files``. Create-input only — not persisted on
the node after upload.
ValidationRule:
properties:
id:
type: string
minLength: 1
title: Id
description: Stable rule id; round-trips through validation results.
name:
anyOf:
- type: string
- type: 'null'
title: Name
description: >-
Optional human-readable rule name. Round-trips through the config
endpoint so renames are preserved.
kind:
type: string
enum:
- ai
- deterministic
title: Kind
description: >-
'deterministic' = a structured check run in code — instant, free,
reproducible; 'ai' = a model judges the rule description against the
extraction. Prefer deterministic whenever the condition is exact and
testable.
default: ai
description:
anyOf:
- type: string
- type: 'null'
title: Description
description: >-
Natural-language description sent to the LLM. Required when kind ==
'ai'.
check:
anyOf:
- oneOf:
- $ref: '#/components/schemas/RangeCheck'
- $ref: '#/components/schemas/DateCheck'
- $ref: '#/components/schemas/ArithmeticCheck'
- $ref: '#/components/schemas/ComparisonCheck'
- $ref: '#/components/schemas/OneOfCheck'
- $ref: '#/components/schemas/RegexCheck'
- $ref: '#/components/schemas/RequiredCheck'
- $ref: '#/components/schemas/ConfidenceCheck'
- $ref: '#/components/schemas/ExpressionCheck'
- $ref: '#/components/schemas/ValidationCheck'
- $ref: '#/components/schemas/AllOfCheck'
- $ref: '#/components/schemas/AnyOfCheck'
discriminator:
propertyName: type
mapping:
all_of:
$ref: '#/components/schemas/AllOfCheck'
any_of:
$ref: '#/components/schemas/AnyOfCheck'
arithmetic:
$ref: '#/components/schemas/ArithmeticCheck'
comparison:
$ref: '#/components/schemas/ComparisonCheck'
confidence:
$ref: '#/components/schemas/ConfidenceCheck'
date:
$ref: '#/components/schemas/DateCheck'
expression:
$ref: '#/components/schemas/ExpressionCheck'
one_of:
$ref: '#/components/schemas/OneOfCheck'
range:
$ref: '#/components/schemas/RangeCheck'
regex:
$ref: '#/components/schemas/RegexCheck'
required:
$ref: '#/components/schemas/RequiredCheck'
validation:
$ref: '#/components/schemas/ValidationCheck'
- type: 'null'
title: Check
description: >-
Structured deterministic check. Required when kind ==
'deterministic'.
severity:
type: string
enum:
- error
- warning
title: Severity
description: >-
How a failed rule is labelled in results. Display-level only — it
never blocks the run.
default: error
source_fields:
items:
type: string
type: array
title: Source Fields
description: Persistent ids of fields this rule references.
additionalProperties: false
type: object
required:
- id
title: ValidationRule
RangeCheck:
properties:
type:
type: string
const: range
title: Type
field:
type: string
minLength: 1
title: Field
description: persistent_id of the bounded numeric field.
min:
anyOf:
- type: number
- type: 'null'
title: Min
max:
anyOf:
- type: number
- type: 'null'
title: Max
additionalProperties: false
type: object
required:
- type
- field
title: RangeCheck
DateCheck:
properties:
type:
type: string
const: date
title: Type
field:
type: string
minLength: 1
title: Field
description: persistent_id of the date field.
earliest:
anyOf:
- type: string
- type: 'null'
title: Earliest
latest:
anyOf:
- type: string
- type: 'null'
title: Latest
additionalProperties: false
type: object
required:
- type
- field
title: DateCheck
ArithmeticCheck:
properties:
type:
type: string
const: arithmetic
title: Type
operands:
items:
type: string
type: array
minItems: 1
title: Operands
description: persistent_ids combined by `operator`.
operator:
type: string
enum:
- sum
- subtract
- product
title: Operator
default: sum
equals:
type: string
minLength: 1
title: Equals
description: persistent_id whose value the result must equal.
tolerance:
type: number
minimum: 0
title: Tolerance
description: Absolute slack allowed around `equals`.
default: 0
additionalProperties: false
type: object
required:
- type
- operands
- equals
title: ArithmeticCheck
ComparisonCheck:
properties:
type:
type: string
const: comparison
title: Type
left:
type: string
minLength: 1
title: Left
description: persistent_id of the left-hand field.
op:
type: string
enum:
- '=='
- '!='
- '>'
- '>='
- <
- <=
title: Op
right:
oneOf:
- $ref: '#/components/schemas/_FieldOperand'
- $ref: '#/components/schemas/_LiteralOperand'
title: Right
discriminator:
propertyName: source
mapping:
field:
$ref: '#/components/schemas/_FieldOperand'
literal:
$ref: '#/components/schemas/_LiteralOperand'
additionalProperties: false
type: object
required:
- type
- left
- op
- right
title: ComparisonCheck
OneOfCheck:
properties:
type:
type: string
const: one_of
title: Type
field:
type: string
minLength: 1
title: Field
allowed:
items:
type: string
type: array
minItems: 1
title: Allowed
case_sensitive:
type: boolean
title: Case Sensitive
default: false
additionalProperties: false
type: object
required:
- type
- field
- allowed
title: OneOfCheck
RegexCheck:
properties:
type:
type: string
const: regex
title: Type
field:
type: string
minLength: 1
title: Field
pattern:
type: string
minLength: 1
title: Pattern
description: Evaluated with google-re2 (linear time); never stdlib re.
additionalProperties: false
type: object
required:
- type
- field
- pattern
title: RegexCheck
RequiredCheck:
properties:
type:
type: string
const: required
title: Type
field:
type: string
minLength: 1
title: Field
additionalProperties: false
type: object
required:
- type
- field
title: RequiredCheck
ConfidenceCheck:
properties:
type:
type: string
const: confidence
title: Type
field:
type: string
minLength: 1
title: Field
description: >-
Field whose extraction confidence to compare. Either the
``persistent_id`` (Studio field-picker path) or the
``sanitized_name`` (SDK/API/assistant path) — the executor resolves
both against the workflow's field tree, same as ``ValidateNode``
deterministic checks.
op:
type: string
enum:
- <
- <=
- '>'
- '>='
- '=='
- '!='
title: Op
threshold:
type: integer
maximum: 100
minimum: 0
title: Threshold
description: Confidence percentile 0..100.
additionalProperties: false
type: object
required:
- type
- field
- op
- threshold
title: ConfidenceCheck
ExpressionCheck:
properties:
type:
type: string
const: expression
title: Type
expression:
type: string
maxLength: 1000
minLength: 1
title: Expression
additionalProperties: false
type: object
required:
- type
- expression
title: ExpressionCheck
description: >-
A CEL expression over the whole extraction, evaluated in a sandbox.
The expression reads one root variable ``data`` and addresses fields by
extracted name (``data.total``, ``data.lineas.map(l, l.importe)``),
never by
``persistent_id`` — so ``field_ids`` returns ``[]`` for it, and a field
rename silently breaks the rule. An expression the evaluator cannot
parse,
cannot answer with a boolean, or prices above its evaluation budget
reports
``inconclusive``.
ADR platform/0036 records why the CEL parser stays out of this package,
so
``expression`` is unvalidated here.
ValidationCheck:
properties:
type:
type: string
const: validation
title: Type
rule_id:
type: string
minLength: 1
title: Rule Id
description: Stable id of the upstream ValidationRule to route on.
status:
type: string
enum:
- fail
- pass
- inconclusive
title: Status
description: >-
Outcome to match. ``fail`` (default) is the common case for routing
to an alert branch.
default: fail
additionalProperties: false
type: object
required:
- type
- rule_id
title: ValidationCheck
description: |-
Routes on the outcome of a named validation rule already run upstream.
Reads from ``state.validation_results`` (populated by an upstream
``ValidateNode``), filtered to the current partition (``split_id``) so
per-partition IfElse routing matches per-partition rule outcomes. The
``status`` values mirror the worker enum
(``anyformat_core.operations.validate_rules`` — ``pass`` / ``fail`` /
``inconclusive``) verbatim: no translation layer, so template tokens
like ``${validation..status}`` render the same three strings
the enum uses.
Unlike every other ``Check`` variant this consumes no extracted fields
(``field_ids`` returns ``[]``); the rule identity is the entire input.
A ``rule_id`` with no matching entry evaluates to ``inconclusive`` — the
rule didn't run for this partition, so the outcome is unknown.
AllOfCheck:
properties:
type:
type: string
const: all_of
title: Type
checks:
items:
oneOf:
- $ref: '#/components/schemas/RangeCheck'
- $ref: '#/components/schemas/DateCheck'
- $ref: '#/components/schemas/ArithmeticCheck'
- $ref: '#/components/schemas/ComparisonCheck'
- $ref: '#/components/schemas/OneOfCheck'
- $ref: '#/components/schemas/RegexCheck'
- $ref: '#/components/schemas/RequiredCheck'
- $ref: '#/components/schemas/ConfidenceCheck'
- $ref: '#/components/schemas/ExpressionCheck'
- $ref: '#/components/schemas/ValidationCheck'
- $ref: '#/components/schemas/AllOfCheck'
- $ref: '#/components/schemas/AnyOfCheck'
discriminator:
propertyName: type
mapping:
all_of:
$ref: '#/components/schemas/AllOfCheck'
any_of:
$ref: '#/components/schemas/AnyOfCheck'
arithmetic:
$ref: '#/components/schemas/ArithmeticCheck'
comparison:
$ref: '#/components/schemas/ComparisonCheck'
confidence:
$ref: '#/components/schemas/ConfidenceCheck'
date:
$ref: '#/components/schemas/DateCheck'
expression:
$ref: '#/components/schemas/ExpressionCheck'
one_of:
$ref: '#/components/schemas/OneOfCheck'
range:
$ref: '#/components/schemas/RangeCheck'
regex:
$ref: '#/components/schemas/RegexCheck'
required:
$ref: '#/components/schemas/RequiredCheck'
validation:
$ref: '#/components/schemas/ValidationCheck'
type: array
maxItems: 32
minItems: 1
title: Checks
additionalProperties: false
type: object
required:
- type
- checks
title: AllOfCheck
description: >-
AND combinator — passes iff every sub-check passes.
Tri-state: FAIL if any child fails; else INCONCLUSIVE if any child is
inconclusive; else PASS. Children are the same closed ``Check`` union
(mutual recursion via forward ref) so combinators nest by construction —
the CheckBuilder UI restricts nesting to a single level for MVP.
``max_length=32`` bounds runaway fan-out per combinator. Practical rules
stay well below the cap; a legitimate use case wanting more should
motivate
the raise rather than land silently. Nesting-depth is bounded by
pydantic's
own recursion during parse.
AnyOfCheck:
properties:
type:
type: string
const: any_of
title: Type
checks:
items:
oneOf:
- $ref: '#/components/schemas/RangeCheck'
- $ref: '#/components/schemas/DateCheck'
- $ref: '#/components/schemas/ArithmeticCheck'
- $ref: '#/components/schemas/ComparisonCheck'
- $ref: '#/components/schemas/OneOfCheck'
- $ref: '#/components/schemas/RegexCheck'
- $ref: '#/components/schemas/RequiredCheck'
- $ref: '#/components/schemas/ConfidenceCheck'
- $ref: '#/components/schemas/ExpressionCheck'
- $ref: '#/components/schemas/ValidationCheck'
- $ref: '#/components/schemas/AllOfCheck'
- $ref: '#/components/schemas/AnyOfCheck'
discriminator:
propertyName: type
mapping:
all_of:
$ref: '#/components/schemas/AllOfCheck'
any_of:
$ref: '#/components/schemas/AnyOfCheck'
arithmetic:
$ref: '#/components/schemas/ArithmeticCheck'
comparison:
$ref: '#/components/schemas/ComparisonCheck'
confidence:
$ref: '#/components/schemas/ConfidenceCheck'
date:
$ref: '#/components/schemas/DateCheck'
expression:
$ref: '#/components/schemas/ExpressionCheck'
one_of:
$ref: '#/components/schemas/OneOfCheck'
range:
$ref: '#/components/schemas/RangeCheck'
regex:
$ref: '#/components/schemas/RegexCheck'
required:
$ref: '#/components/schemas/RequiredCheck'
validation:
$ref: '#/components/schemas/ValidationCheck'
type: array
maxItems: 32
minItems: 1
title: Checks
additionalProperties: false
type: object
required:
- type
- checks
title: AnyOfCheck
description: |-
OR combinator — passes iff any sub-check passes.
Tri-state: PASS if any child passes; else INCONCLUSIVE if any child is
inconclusive; else FAIL.
``max_length=32`` — see :class:`AllOfCheck` for the rationale.
StringField:
properties:
persistent_id:
anyOf:
- type: string
- type: 'null'
title: Persistent Id
description: >-
Server-assigned stable field identity (UUID), emitted by GET
/definition/. Echo it back on updates — including across renames —
to keep quality metrics, ground-truth links and analytics attached
to the field. Matching is scoped to the field's nesting level in the
immediately-prior version; an id that matches nothing at its level,
appears on more than one field, or is sent at workflow creation is
rejected with 400. Omit for new fields; when omitted on an update,
the server matches the field to the prior version by name.
name:
type: string
minLength: 1
title: Name
description: Field name. Used as the key in the extraction response.
description:
type: string
title: Description
description: Free-form description shown to the extraction model.
default: ''
source:
type: string
enum:
- extraction
- smart_lookup
- lookup_if_missing
title: Source
description: >-
How the field gets its value: extraction (from the document),
smart_lookup (from the lookup file, always overwrites) or
lookup_if_missing (extracted too; the lookup only fills it where
extraction left no value).
default: extraction
data_type:
type: string
const: string
title: Data Type
additionalProperties: false
type: object
required:
- name
- data_type
title: StringField
IntegerField:
properties:
persistent_id:
anyOf:
- type: string
- type: 'null'
title: Persistent Id
description: >-
Server-assigned stable field identity (UUID), emitted by GET
/definition/. Echo it back on updates — including across renames —
to keep quality metrics, ground-truth links and analytics attached
to the field. Matching is scoped to the field's nesting level in the
immediately-prior version; an id that matches nothing at its level,
appears on more than one field, or is sent at workflow creation is
rejected with 400. Omit for new fields; when omitted on an update,
the server matches the field to the prior version by name.
name:
type: string
minLength: 1
title: Name
description: Field name. Used as the key in the extraction response.
description:
type: string
title: Description
description: Free-form description shown to the extraction model.
default: ''
source:
type: string
enum:
- extraction
- smart_lookup
- lookup_if_missing
title: Source
description: >-
How the field gets its value: extraction (from the document),
smart_lookup (from the lookup file, always overwrites) or
lookup_if_missing (extracted too; the lookup only fills it where
extraction left no value).
default: extraction
data_type:
type: string
const: integer
title: Data Type
additionalProperties: false
type: object
required:
- name
- data_type
title: IntegerField
FloatField:
properties:
persistent_id:
anyOf:
- type: string
- type: 'null'
title: Persistent Id
description: >-
Server-assigned stable field identity (UUID), emitted by GET
/definition/. Echo it back on updates — including across renames —
to keep quality metrics, ground-truth links and analytics attached
to the field. Matching is scoped to the field's nesting level in the
immediately-prior version; an id that matches nothing at its level,
appears on more than one field, or is sent at workflow creation is
rejected with 400. Omit for new fields; when omitted on an update,
the server matches the field to the prior version by name.
name:
type: string
minLength: 1
title: Name
description: Field name. Used as the key in the extraction response.
description:
type: string
title: Description
description: Free-form description shown to the extraction model.
default: ''
source:
type: string
enum:
- extraction
- smart_lookup
- lookup_if_missing
title: Source
description: >-
How the field gets its value: extraction (from the document),
smart_lookup (from the lookup file, always overwrites) or
lookup_if_missing (extracted too; the lookup only fills it where
extraction left no value).
default: extraction
data_type:
type: string
const: float
title: Data Type
additionalProperties: false
type: object
required:
- name
- data_type
title: FloatField
BooleanField:
properties:
persistent_id:
anyOf:
- type: string
- type: 'null'
title: Persistent Id
description: >-
Server-assigned stable field identity (UUID), emitted by GET
/definition/. Echo it back on updates — including across renames —
to keep quality metrics, ground-truth links and analytics attached
to the field. Matching is scoped to the field's nesting level in the
immediately-prior version; an id that matches nothing at its level,
appears on more than one field, or is sent at workflow creation is
rejected with 400. Omit for new fields; when omitted on an update,
the server matches the field to the prior version by name.
name:
type: string
minLength: 1
title: Name
description: Field name. Used as the key in the extraction response.
description:
type: string
title: Description
description: Free-form description shown to the extraction model.
default: ''
source:
type: string
enum:
- extraction
- smart_lookup
- lookup_if_missing
title: Source
description: >-
How the field gets its value: extraction (from the document),
smart_lookup (from the lookup file, always overwrites) or
lookup_if_missing (extracted too; the lookup only fills it where
extraction left no value).
default: extraction
data_type:
type: string
const: boolean
title: Data Type
additionalProperties: false
type: object
required:
- name
- data_type
title: BooleanField
DateField:
properties:
persistent_id:
anyOf:
- type: string
- type: 'null'
title: Persistent Id
description: >-
Server-assigned stable field identity (UUID), emitted by GET
/definition/. Echo it back on updates — including across renames —
to keep quality metrics, ground-truth links and analytics attached
to the field. Matching is scoped to the field's nesting level in the
immediately-prior version; an id that matches nothing at its level,
appears on more than one field, or is sent at workflow creation is
rejected with 400. Omit for new fields; when omitted on an update,
the server matches the field to the prior version by name.
name:
type: string
minLength: 1
title: Name
description: Field name. Used as the key in the extraction response.
description:
type: string
title: Description
description: Free-form description shown to the extraction model.
default: ''
source:
type: string
enum:
- extraction
- smart_lookup
- lookup_if_missing
title: Source
description: >-
How the field gets its value: extraction (from the document),
smart_lookup (from the lookup file, always overwrites) or
lookup_if_missing (extracted too; the lookup only fills it where
extraction left no value).
default: extraction
data_type:
type: string
const: date
title: Data Type
additionalProperties: false
type: object
required:
- name
- data_type
title: DateField
DatetimeField:
properties:
persistent_id:
anyOf:
- type: string
- type: 'null'
title: Persistent Id
description: >-
Server-assigned stable field identity (UUID), emitted by GET
/definition/. Echo it back on updates — including across renames —
to keep quality metrics, ground-truth links and analytics attached
to the field. Matching is scoped to the field's nesting level in the
immediately-prior version; an id that matches nothing at its level,
appears on more than one field, or is sent at workflow creation is
rejected with 400. Omit for new fields; when omitted on an update,
the server matches the field to the prior version by name.
name:
type: string
minLength: 1
title: Name
description: Field name. Used as the key in the extraction response.
description:
type: string
title: Description
description: Free-form description shown to the extraction model.
default: ''
source:
type: string
enum:
- extraction
- smart_lookup
- lookup_if_missing
title: Source
description: >-
How the field gets its value: extraction (from the document),
smart_lookup (from the lookup file, always overwrites) or
lookup_if_missing (extracted too; the lookup only fills it where
extraction left no value).
default: extraction
data_type:
type: string
const: datetime
title: Data Type
additionalProperties: false
type: object
required:
- name
- data_type
title: DatetimeField
EnumField:
properties:
persistent_id:
anyOf:
- type: string
- type: 'null'
title: Persistent Id
description: >-
Server-assigned stable field identity (UUID), emitted by GET
/definition/. Echo it back on updates — including across renames —
to keep quality metrics, ground-truth links and analytics attached
to the field. Matching is scoped to the field's nesting level in the
immediately-prior version; an id that matches nothing at its level,
appears on more than one field, or is sent at workflow creation is
rejected with 400. Omit for new fields; when omitted on an update,
the server matches the field to the prior version by name.
name:
type: string
minLength: 1
title: Name
description: Field name. Used as the key in the extraction response.
description:
type: string
title: Description
description: Free-form description shown to the extraction model.
default: ''
source:
type: string
enum:
- extraction
- smart_lookup
- lookup_if_missing
title: Source
description: >-
How the field gets its value: extraction (from the document),
smart_lookup (from the lookup file, always overwrites) or
lookup_if_missing (extracted too; the lookup only fills it where
extraction left no value).
default: extraction
data_type:
type: string
const: enum
title: Data Type
enum_options:
items:
$ref: '#/components/schemas/EnumOption'
type: array
minItems: 1
title: Enum Options
additionalProperties: false
type: object
required:
- name
- data_type
- enum_options
title: EnumField
MultiSelectField:
properties:
persistent_id:
anyOf:
- type: string
- type: 'null'
title: Persistent Id
description: >-
Server-assigned stable field identity (UUID), emitted by GET
/definition/. Echo it back on updates — including across renames —
to keep quality metrics, ground-truth links and analytics attached
to the field. Matching is scoped to the field's nesting level in the
immediately-prior version; an id that matches nothing at its level,
appears on more than one field, or is sent at workflow creation is
rejected with 400. Omit for new fields; when omitted on an update,
the server matches the field to the prior version by name.
name:
type: string
minLength: 1
title: Name
description: Field name. Used as the key in the extraction response.
description:
type: string
title: Description
description: Free-form description shown to the extraction model.
default: ''
source:
type: string
enum:
- extraction
- smart_lookup
- lookup_if_missing
title: Source
description: >-
How the field gets its value: extraction (from the document),
smart_lookup (from the lookup file, always overwrites) or
lookup_if_missing (extracted too; the lookup only fills it where
extraction left no value).
default: extraction
data_type:
type: string
const: multi_select
title: Data Type
enum_options:
items:
$ref: '#/components/schemas/EnumOption'
type: array
minItems: 1
title: Enum Options
additionalProperties: false
type: object
required:
- name
- data_type
- enum_options
title: MultiSelectField
ObjectField:
properties:
persistent_id:
anyOf:
- type: string
- type: 'null'
title: Persistent Id
description: >-
Server-assigned stable field identity (UUID), emitted by GET
/definition/. Echo it back on updates — including across renames —
to keep quality metrics, ground-truth links and analytics attached
to the field. Matching is scoped to the field's nesting level in the
immediately-prior version; an id that matches nothing at its level,
appears on more than one field, or is sent at workflow creation is
rejected with 400. Omit for new fields; when omitted on an update,
the server matches the field to the prior version by name.
name:
type: string
minLength: 1
title: Name
description: Field name. Used as the key in the extraction response.
description:
type: string
title: Description
description: Free-form description shown to the extraction model.
default: ''
source:
type: string
enum:
- extraction
- smart_lookup
- lookup_if_missing
title: Source
description: >-
How the field gets its value: extraction (from the document),
smart_lookup (from the lookup file, always overwrites) or
lookup_if_missing (extracted too; the lookup only fills it where
extraction left no value).
default: extraction
data_type:
type: string
const: object
title: Data Type
nested_fields:
items:
oneOf:
- $ref: '#/components/schemas/StringField'
- $ref: '#/components/schemas/IntegerField'
- $ref: '#/components/schemas/FloatField'
- $ref: '#/components/schemas/BooleanField'
- $ref: '#/components/schemas/DateField'
- $ref: '#/components/schemas/DatetimeField'
- $ref: '#/components/schemas/EnumField'
- $ref: '#/components/schemas/MultiSelectField'
discriminator:
propertyName: data_type
mapping:
boolean:
$ref: '#/components/schemas/BooleanField'
date:
$ref: '#/components/schemas/DateField'
datetime:
$ref: '#/components/schemas/DatetimeField'
enum:
$ref: '#/components/schemas/EnumField'
float:
$ref: '#/components/schemas/FloatField'
integer:
$ref: '#/components/schemas/IntegerField'
multi_select:
$ref: '#/components/schemas/MultiSelectField'
string:
$ref: '#/components/schemas/StringField'
type: array
minItems: 1
title: Nested Fields
description: >-
Child fields of this object. Only one level of nesting is supported
— an object field's children must be non-object types.
additionalProperties: false
type: object
required:
- name
- data_type
- nested_fields
title: ObjectField
_FieldOperand:
properties:
source:
type: string
const: field
title: Source
field:
type: string
minLength: 1
title: Field
description: persistent_id of the compared field.
additionalProperties: false
type: object
required:
- source
- field
title: _FieldOperand
_LiteralOperand:
properties:
source:
type: string
const: literal
title: Source
value:
anyOf:
- type: number
- type: string
- type: boolean
title: Value
additionalProperties: false
type: object
required:
- source
- value
title: _LiteralOperand
EnumOption:
properties:
name:
type: string
minLength: 1
title: Name
description:
type: string
minLength: 1
title: Description
description: Free-form description shown to the model.
additionalProperties: false
type: object
required:
- name
- description
title: EnumOption
securitySchemes:
ApiKeyAuth:
type: http
description: >-
API key issued from app.anyformat.ai/api-key. Send as `Authorization:
Bearer `.
scheme: bearer
````