How to Use Mistral document text extraction in Factory Droid
Mistral OCR 3 extracts text, tables, and embedded images from PDFs and images into page-based Markdown. Use it to turn scanned reports and complex tables into material an agent can search, compare, or analyze.
Prompts to try with Mistral document text extraction in Factory Droid
Copy an example into Factory Droid, or ask in your own words.
Use Mistral document text extraction on https://arxiv.org/pdf/2201.04234 and save the extracted page text as Markdown in this project's research directory.
Use Mistral document text extraction on Microsoft's 2024 annual report PDF and build a CSV from the extracted income statement, preserving its headings.
Use Mistral document text extraction on the Mistral 7B paper at https://arxiv.org/pdf/2310.06825 and add its extracted text to this project's search index.
Use Mistral document text extraction on Apple's 2023 and 2024 annual report PDFs and write a script that compares their extracted segment tables.
Use Mistral document text extraction on https://arxiv.org/pdf/2201.04234 and build a page-by-page reader in this project using the returned text.
How to set up Mistral document text extraction in Factory Droid
Pick the way that suits you. Each takes about two minutes.
Use Spiral's MCP server
- Add the server
Droid saves it to your user config, ~/.factory/mcp.json, so it works in every project.
droid mcp add spiral https://api.spiral.ai/mcp --type http
- Sign in
Start droid, type /mcp, select spiral and authenticate it. Sign in to Spiral in the browser and click Authorize. No API key.
- Check it
Spiral shows as connected. needs authentication means the sign-in did not finish: run /mcp again.
droid mcp list
Use the Spiral skill
- Send Factory Droid this message
Your agent reads Spiral's skill and sets itself up.
set up https://spiral.ai/SKILL.md
- Sign in or paste a key
Approve Spiral in the browser, or give the agent a key from the console.
Use the Spiral CLI
- Install and sign in
Installs the Spiral CLI, signs you in and adds its skill to your coding agents.
npx -y spiral-cli@latest init
Call the Spiral API
- Create an API key
In the console, under API keys.
- Run Mistral document text extraction
One request; the result comes back by polling or a webhook.
curl https://api.spiral.ai/v1/runs \ -H "Authorization: Bearer $SPIRAL_API_KEY" \ -H "Content-Type: application/json" \ -H "Idempotency-Key: $(uuidgen)" \ -d '{ "provider": "mistral", "model": "mistral-ocr-3-0", "input": { "document": { "document_url": "<document_url>" } } }'
Then paste one of the prompts above: Factory Droid finds Mistral document text extraction on Spiral, shows the price and runs it.
Mistral document text extraction providers and prices
Spiral runs Mistral document text extraction through Mistral. Open a card for today's price, with every fee included.
Settings you can ask for
Mistral document text extraction works from your document annotation prompt. Ask Factory Droid for any of these in your prompt.
| Setting | Allowed values | Default | Description |
|---|---|---|---|
| Image limit | a whole number | Max images to extract | |
| Table format | markdown or html | ||
| Extract footer | on or off | off | Extract the page footer into the response's |
| Extract header | on or off | off | Extract the page header into the response's |
| Image min size | a whole number | Minimum height and width of image to extract | |
| Include blocks | on or off | on | Return paragraph-level bounding boxes for all content blocks in the response |
| Include image base64 | on or off | Include image URLs in response | |
| Confidence scores granularity | word, page, or block | Granularity for confidence scores: 'page' (aggregate only), 'word' (per-word scores), or 'block' (per-block scores). Defaults to None (no confidence scores) to keep response payload small. |
Every API field of Mistral OCR 3.0 is on its tool page: open its card above.
What can you do with Mistral document text extraction in Factory Droid?
Build a searchable collection of scanned reports using Mistral's extracted page text and images. A coding agent can add the results to a project's document index or write a table comparison tool.
Search scanned business archives
Extract page text from archived company reports, then index it so staff can find passages without reading every scan.
Compare annual report tables
Extract financial tables from successive public annual reports and compare rows while retaining their column headings.
Prepare research papers for retrieval
Convert technical PDFs into page-based text for a research search tool, keeping extracted figures available alongside the text.
Extract operational document fields
Turn authorized supplier invoices into structured records for a reviewer to check before entering them into accounting software.
Compare manual revisions
Extract text and tables from two editions of a product manual, then identify changed specifications and instructions.
Make historical publications searchable
Transcribe scanned institutional publications into searchable text, including handwritten annotations where present.
Catalog figures in technical reports
Extract embedded figures and request captions to help readers locate relevant diagrams within a report collection.
More on Mistral document text extraction and Factory Droid
Overview
Mistral document text extraction in other agents
Questions about Mistral document text extraction in Factory Droid
Can Factory Droid use Mistral document text extraction?
Yes. Add Spiral to Factory Droid as an MCP server, and Factory Droid can find and run Mistral document text extraction like any built-in tool.
How much does Mistral document text extraction cost in Factory Droid?
$0.0026 per page ($0.0039 with document or bounding-box annotations). You pay from your Spiral balance as you go, with no subscription, and a failed run is not charged. Your Factory Droid plan is separate.
Do I need a Mistral account to use Mistral document text extraction in Factory Droid?
No. Spiral runs Mistral document text extraction on its own Mistral account. You sign in to Spiral once and never paste a key.
How long does it take to set up Mistral document text extraction in Factory Droid?
About two minutes: add the server, then sign in, then check it.
Does Factory Droid wait while Mistral document text extraction runs?
Mistral document text extraction runs in the background. Factory Droid starts the run, checks on it and gives you the json when it's ready.
Where does the extracted information come from?
It comes from the PDF or image submitted for processing. A report's contents therefore reflect that document's edition, including any outdated figures it contains.
How large a document can Mistral OCR process?
Mistral's chunking cookbook lists typical limits of approximately 50 MB and 1,000 pages per processing call. It recommends splitting larger documents into smaller parts and combining their results.
Can OCR 3 extract tables with merged cells?
Yes. Mistral documents support for merged cells, multi-row blocks, table headers, and column hierarchies.
Does OCR 3 handle handwriting?
Mistral describes support for cursive, mixed-content annotations, and handwriting layered over printed forms. These are supported document types, rather than a published guarantee of perfect transcription.
Does OCR 3 return paragraph bounding boxes?
Mistral's current documentation says paragraph-level block extraction requires OCR 4 or newer. Do not build an OCR 3 workflow that depends on those paragraph coordinates.
Start using Mistral document text extraction in Factory Droid
One Spiral account for every tool your agent needs. No subscription; you pay per run.
Paste this into Factory Droid:
set up https://spiral.ai/SKILL.md


