Skip to main content
POST
/
v1
/
agentic
/
documents
/
induce_schema
Induce Schema
curl --request POST \
  --url https://api.example.com/v1/agentic/documents/induce_schema \
  --header 'Content-Type: application/json' \
  --data '
{
  "documents": [
    {
      "document_id": "<string>",
      "upload_key": "<string>"
    }
  ],
  "use_case": "<string>",
  "skill": "<string>",
  "model_id": "<string>"
}
'
{
  "job_id": "3c90c3cc-0d44-4b50-8888-8dd25736052a",
  "message": "<string>",
  "schema_id": "<string>"
}

Headers

​
authorization
string | null
​
refresh-token
string | null

Body

application/json

POST /v1/agentic/documents/induce_schema -- async schema induction job.

Runs the agent over the supplied document sample to produce a frozen :class:Schema. Submit, poll GET /v1/models/{model_id}/status until status="completed", then fetch the result via GET /v1/agentic/documents/schemas/{model_id}.

​
documents
DocumentRef · object[]
required

Sample of uploaded documents the agent inspects to draft the schema.

Minimum array length: 1
​
use_case
string
required

Free-text description of what the resulting features will be used for.

​
skill
string
required

Skill slug; one of the values returned by GET /v1/agentic/documents/skills.

​
model_id
string | null

Custom model_id for the resulting schema; auto-generated if omitted.

Response

Successful Response

POST /v1/agentic/documents/induce_schema -- async submission response.

job_id is the canonical job handle (poll GET /v1/results/{job_id}). schema_id is the id of the schema being produced — used to fetch the typed result via GET /v1/agentic/documents/schemas/{schema_id} once the job completes.

​
job_id
required

Server-assigned id of the submitted job

​
status
enum<string>
required
Available options:
pending,
running,
completed,
failed
​
message
string
required
​
schema_id
string
required