Document Parse | BetterAI Docs

#Document Parse - `/parse-content/document-parse`

POST/parse-content/document-parse

Receives a file and a JSON schema to extract and structure its content via LLM. Use it when you need to turn unstructured documents (PDFs, Word, text) into organized data ready for consumption.

#Parameters

Parameter	Type	Description	Example
`metadata`	`string` (JSON)	Auxiliary metadata of the document, free for the client to use.	`{"client": "Acme", "doc_type": "contract"}`
`document_schema`	`string` (JSON)	Schema that defines the fields to extract. Each key takes `type` and `description`.	`{"summary": {"type": "str", "description": "Resumo do documento"}}`
`file`	`file` (upload)	File to be processed. Accepted formats: `txt`, `md`, `pdf`, `docx`. Limit: 50MB.	`contrato.pdf`
`config`	`string` (JSON) — optional	Parser settings: LLM model, custom instructions and debug mode.	`{"model_provider": "OpenAI", "model_id": "gpt-4.1-mini", "debug_mode": false}`

Note: the job_id is generated automatically by the server and returned in the response.

Schema formats: https://github.com/enzoschitini/better-ai/blob/production/doc/Modules/Content%20Parse/Json%20To%20Pydantic.md

#Config (explanation of each parameter)

The config field is optional. When not sent, the API uses the parser's internal default_config.

Field	Type	What it does	Default	When to adjust
`model_provider`	`string`	Defines the language model provider. Currently the main flows use `OpenAI` and `Groq`.	`"OpenAI"`	Change it when you want to switch provider because of cost, latency or availability.
`model_id`	`string`	Identifies the model used for extraction (`gpt-4.1-mini`, for example).	`"gpt-4.1-mini"`	Adjust it when you need higher quality, lower cost or a different context window.
`max_input_tokens`	`integer`	Limit of input tokens allowed to validate the size of the context before execution.	`1000000`	Lower it to impose a cost limit or raise it in scenarios with very long documents.
`debug_mode`	`boolean`	Turns on logs and debugging information of the parsing agent.	`false`	Use `true` for troubleshooting and go back to `false` in production.
`instructions`	`string`	Operational instructions on how to extract the data (style, rules, focus).	`"Extraia dados do texto"`	Customize it to guide the model toward formats and criteria specific to your domain.
`description`	`string`	Role/expected behavior of the agent, reinforcing the response format and the fallback with `null`.	`"Leia o texto e extraia as informações relevantes conforme o esquema definido..."`	Adjust it for business scenarios with their own filling and validation policies.

Complete config example:

{
  "model_provider": "OpenAI",
  "model_id": "gpt-4.1-mini",
  "max_input_tokens": 1000000,
  "debug_mode": false,
  "instructions": "Extraia dados do texto",
  "description": "Leia o texto e extraia as informações relevantes conforme o esquema definido. Retorne um JSON estruturado com os dados extraídos. Caso não encontre alguma informação, retorne null para aquele campo."
}

#Request

curl --location 'http://localhost:8000/parse-content/document-parse' \
--header 'X-API-Key: ******' \
--form 'metadata="{\"value1\": \"value3\"}"' \
--form 'document_schema="{
  \"summary\": {
    \"type\": \"str\",
    \"description\": \"Resumo do conteúdo do arquivo\"
  }
}"' \
--form 'config="{
  \"model_provider\": \"OpenAI\",
  \"model_id\": \"gpt-4.1-mini\",
  \"debug_mode\": true,
  \"instructions\": \"Extraia dados do texto\",
  \"description\": \"Leia o texto e extraia as informações relevantes conforme o esquema definido. Retorne um JSON estruturado com os dados extraídos. Caso não encontre alguma informação, retorne null para aquele campo.\"
}"' \
--form 'file=@"/path/to/file"'

#Response

{
    "job_id": "job_1788198742619260200fKBY",
    "status": "success",
    "status_code": 200,
    "content": {
        "summary": "..."
    },
    "time": {
        "start": "2026-08-31 14:52:22",
        "end": "2026-08-31 14:52:30",
        "duration_seconds": 8.0
    }
}

#Document Parse - /parse-content/document-parse

#Parameters

#Config (explanation of each parameter)

#Request

#Response

#Document Parse - `/parse-content/document-parse`