Structured Extraction with Instructor and Brazil CVM Filings

Using Instructor CVM filings workflows? Pull filings from apicvm, then run Instructor schemas on downloaded or extracted text. apicvm is not an Instructor plugin.

The problem

Structured LLM pipelines that cite Brazilian filings need durable document UUIDs before schema validation. Scraped HTML is a poor grounding source.

Solution flow

list documents → download/extract → Instructor schema → cite document_id

Fetch then structure

import os, requests

BASE = os.environ["APICVM_URL"]
H = {"Authorization": f"Bearer {os.environ['APICVM_KEY']}"}

docs = requests.get(
    f"{BASE}/v1/documents",
    headers=H,
    params={"ticker": "PETR4", "type": "FRE", "year": 2024, "name": "Risco", "perPage": 20},
).json()["data"]
# Pass docs[i]["id"] into your download/extract step, then Instructor.create(...)

Do not invent a sync "get full FRE text" call — extraction is async via callback (Pro).

Current limitations

Student keys cannot run text extraction (403)
Keep document UUIDs in every structured record for auditability
Empty name filters can still return corpus gaps

Next steps

Ready to integrate?

Get an API key and start querying Brazilian CVM filings programmatically.

Read the API docs Get API key