fastocr-sdk 0.3.0

Python client for the FastOCR Document OCR API. Send a scanned PDF or an image, get back its text or a searchable PDF.

Website · Docs · Get an API key

Install

pip install fastocr-sdk
export FASTOCR_API_KEY="your_api_key"

Python 3.9 or newer. The distribution is fastocr-sdk and the import is fastocr_sdk. Do not install the unrelated fastocr package from PyPI.

Quickstart

from fastocr_sdk import FastOCR

with FastOCR() as client:
    print(client.extract_text("scanned_contract.pdf"))

extract_text uploads the file, runs OCR, waits for it to finish, and returns the text. Images work the same way; pass filename when you send bytes.

What you get back

Examples

Runnable scripts are in examples/, with RAG, archiving and cURL recipes in examples/cookbooks/. Full reference →

Searchable PDF

with FastOCR() as client:
    job = client.documents.process("scanned_archive.pdf")
    client.documents.download_searchable_pdf(job.id, "searchable.pdf")

Async and batches

import asyncio
from fastocr_sdk import AsyncFastOCR

async def main():
    async with AsyncFastOCR() as client:
        print(await client.extract_text("invoice.pdf"))

asyncio.run(main())

Accounts process 2 documents at once by default. Cap concurrency with a semaphore, as in examples/async_batch.py.

Errors

Every error is a FastOCRError with status_code, code, request_id and message. Branch on code, never on message.

from fastocr_sdk import DocumentFailedError, FastOCR, FastOCRError

with FastOCR() as client:
    try:
        text = client.extract_text("scan.pdf")
    except DocumentFailedError as error:
        print(error.code, error.retryable, error.document_id)
    except FastOCRError as error:
        print(error.status_code, error.code, error.request_id)

If the account runs out of pages partway, extract_text raises PartialResultError instead of returning truncated text; pass allow_partial=True to accept it. Transient failures are retried and a document is never created twice.

Full reference: retries, idempotency, out-of-pages recovery, configuration.

RAG integrations

Use langchain-fastocr for LangChain or llama-index-readers-fastocr for LlamaIndex. Both are built on this SDK.

Languages and files

• 100+ languages, detected automatically. No language setting is needed.
• PDF files up to 1 GiB, and PNG, JPG, JPEG, GIF, WebP and BMP images up to 20 MiB. An image counts as one page.
• Sync (FastOCR) and async (AsyncFastOCR) clients. One runtime dependency: httpx.

Pages and pricing

Any signed-in FastOCR account can create an API key and start right away. Create one in the developer console; you accept the API terms with your first key. Every account gets 50 free pages a month, shared with PDFs on the website, and an image sent to the API is one page. When you need more, buy a Page Pack on the pricing page. Packs and the page balance are the same on the website and in the API. Check what is left with client.get_balance().

Support and license

Questions: support@fastocr.org. Bugs and requests: GitHub issues. Full reference: docs/reference.md. MIT license; see LICENSE.