Integration · Documents

Connect Claude & ChatGPT to Mindee OCR

Ask your AI to pull the total, date and supplier out of an invoice, sort a stack of documents by type, or split a PDF that holds several invoices. It sends the document to one of your Mindee models, follows the job and reads back the result. Data from documents lands in your chat, ready to check or pass on.

Connect Mindee OCR free

Free plan, no credit card. Takes about a minute.

Try asking

“Which Mindee models do I have?”
The AI lists your models, such as invoice, receipt or custom extraction models, with the IDs needed to use them.
“Extract the supplier, date and total from the invoice at this link.”
It sends the file's public URL to your invoice model, checks the job until it is done and reads the extracted fields.
“What kind of document is this?”
It sends the document to your classification model and reads back the result.

What your AI can do

Extraction

Send a document by public URL or as a file to an extraction model and read the extracted fields.

Classify, crop and split

Detect a document's type, find several documents in one image, or split a multi-document PDF into page ranges.

Models and jobs

List your models to find the right model ID, and follow each job until the result is ready.

“This photo has four receipts on it. Find each one.”
It sends the image to your crop model and returns where each receipt sits in the photo.
“This PDF has several invoices in it. Where does each one start?”
It sends the file to your split model and returns the page ranges of each document.
“Include the raw text too.”
It sends the extraction again with raw_text switched on, so the result includes the document's full text.

10 tools for Mindee OCR

These tools are switched on when you connect. You can switch any of them off, or require your approval before it runs.

classify_documentChanges dataSend a document to a Mindee classification model to detect what kind of document it is. Pass `model_id` and a public HTTPS `url` (or `file_base64`). Returns a job ID to check with get_job_status.
crop_documentChanges dataSend an image to a Mindee crop model to detect each document in it (e.g. several receipts in one photo). Pass `model_id` and a public HTTPS `url` (or `file_base64`). Returns a job ID.
extract_documentChanges dataSend a document to a Mindee extraction model to pull out its data (e.g. invoice totals, dates, supplier). Pass `model_id` and the file as a public HTTPS `url` (or `file_base64`). Optional: `confidence`, `raw_text`, `polygon`, `text_context`. Returns a job ID; check it with get_job_status, then fetch the result.
get_classification_resultGet the result of a completed classification by `inference_id`.
get_crop_resultGet the result of a completed crop by `inference_id`: the location of each document detected.
get_extraction_resultGet the extracted fields of a completed extraction by `inference_id`.
get_job_statusCheck the processing status of a submitted document by its job ID (pass it as `project_id`). When done it points to the result's inference ID.
get_split_resultGet the result of a completed split by `inference_id`: the page ranges of each document found.
list_modelsList your Mindee models (e.g. invoice, receipt or custom extraction models), filtered by `name` or `model_type`. Use to find the `model_id` needed to process a document.
split_documentChanges dataSend a multi-document file (e.g. a PDF with several invoices) to a Mindee split model to find where each document starts and ends. Pass `model_id` and a public HTTPS `url` (or `file_base64`). Returns a job ID.

Set up in three steps

1
Pick the app
Create a free PipMCP account and choose Mindee OCR from the app list.
2
Paste your key
Paste a Mindee API key. Log in to app.mindee.com, click Settings in the left menu and open the API Keys tab. Click + Create API Key, name it and copy the key, as it cannot be shown again. Your model IDs are listed in the Mindee platform, and the AI can also look them up.
3
Add the link to your AI
You get a personal MCP link. Add it to Claude, ChatGPT or Cursor:
1. Click your name, then Settings › Connectors › Add custom connector.
2. Paste your link as the Remote MCP server URL.
3. Switch it on from the + menu in a chat.

Questions

What can the AI do in Mindee?

The AI can list your Mindee models, send documents to extraction, classification, crop and split models, check job status and read the results. It works with the models you have set up in Mindee.

Does the AI see my Mindee credentials?

No. Your Mindee credentials are encrypted at rest and never shown to the AI. After you save them, they are not shown again, not even to you. The AI only sees the results of the tools it calls.

Can I control what the AI is allowed to do?

Yes. You choose which tools are switched on, for example extraction only. Each document you send is processed by Mindee under your account, and you can require your approval before a document is sent. Every tool call is logged.

Does it work with ChatGPT?

Yes. In ChatGPT, open Settings › Apps & Connectors › Advanced, turn on Developer mode and add your PipMCP link. Developer mode needs a paid ChatGPT plan: Plus, Pro, Business or Enterprise. The same link also works in Claude (Settings › Connectors › Add custom connector), Cursor and other MCP clients.

What does it cost?

PipMCP has a free plan with no credit card required. Paid plans bill per completed task. You also need your own Mindee account.

How does the AI send a document to Mindee?

The AI passes a public HTTPS link to the file, or the file itself encoded as base64, together with the ID of one of your models. Mindee returns a job ID, and the AI checks the job until it is done and then fetches the result.