How to Connect OCR.space to ChatGPT via MCP

💡 No install? Use cloud.anythingmcp.com directly. Sign in, click Connectors → OCR.space, paste your credentials, mint an MCP API key and you're done. No Docker, no git clone, no local server.

OCR.space + ChatGPT

OCR.space OCR API: read the text of images and PDFs (receipts, invoices, scans, screenshots, photos of documents) from a public URL, with automatic language detection, table mode for receipts and a searchable-PDF export. Free API key.

Prerequisites

See the full setup instructions baked into the connector (visible in the in-app store when you select the connector). The required environment variables for this connector are:

OCR_SPACE_API_KEY

Step 1 — Get credentials

Setting up takes about 1 minute. You need an OCR.space API key: a free key, which arrives by e-mail, or the key of a PRO plan.

1. Request a free API key at ocr.space/ocrapi/freekey. OCR.space sends it to you by e-mail.
2. Paste only the key into OCR_SPACE_API_KEY, without spaces or quotes. The connector sends it in the apikey header.
3. Install the connector. The install check reads a small sample image, which counts as one request.
4. Run ocr_space_read_url on a small public image, for example a screenshot or a photo of a receipt: it returns the text of every page.

What it reads: the text of images (JPG, PNG, GIF, TIF, BMP, WEBP) and PDFs at a public URL: receipts, invoices, delivery notes, scanned letters, screenshots and photos of documents. ocr_space_read_url returns one text per page, with line breaks, plus exit codes. ocr_space_make_searchable_pdf turns a scan into a searchable PDF (an invisible text layer over the original) and returns its download link together with the recognised text.

The file must be public: OCR.space downloads the file itself, so the URL must open from the internet without a login. A link to a private drive or to an e-mail attachment fails with E217 (could not download). When the URL has no file extension, the AI tells OCR.space the file type.

Engines and languages: engine 2 (the default) is fast, detects the language itself and handles numbers and special characters well. Engine 3 is the most accurate, with 200+ languages, handwriting and complex layouts, but it is slower. Language auto works with both; when it picks the wrong language, the AI sets a code such as ger, fre, ita, spa, chs or jpn.

Receipts, small and rotated images: table mode keeps receipts and tables line by line in reading order. Upscaling helps with small or blurry images, scans and screenshots. Orientation detection turns rotated photos the right way round before reading them.

Searchable PDFs: the download link (SearchablePDFURL) expires after about an hour, so save the file soon. On the free plan the PDF carries a watermark. The text layer can be shown highlighted or kept invisible in PDF viewers.

Read-only: both tools only read, although they are POST requests. Nothing changes in your OCR.space account, so there is nothing to confirm and no tool is switched off at install.

Limits: on the free plan a file can be up to 1 MB and a PDF up to 3 pages, with 25,000 requests a month and 500 a day per IP address. Engine 3 is limited to 1,000 pages a month on the free plan. PRO keys allow 5 MB per file and more pages.

Reading the answer: OCRExitCode 1 means all pages were read, 2 some, 3 or 4 none, and ErrorMessage says why. OCR.space reports processing errors with HTTP 200; the connector turns them into a failed call carrying OCR.space's message.

Common errors: 403 E555 means the API key is wrong. 400 E400 means no file was given, E204 a wrong parameter. E217 is a URL that cannot be downloaded. 429 or E500 means the monthly or daily limit is reached: try again later or use a PRO key.

The connector does not upload files from your computer and cannot read private drive links or e-mail attachments: it only works with a public URL.

Step 2 — Install the adapter

curl -fsSL https://raw.githubusercontent.com/HelpCode-ai/anythingmcp/main/docker-compose.quickstart.yml -o docker-compose.yml
printf 'JWT_SECRET=%s\nENCRYPTION_KEY=%s\n' "$(openssl rand -hex 32)" "$(openssl rand -hex 32)" > .env
docker compose up -d

Step 3 — Add the connector in ChatGPT

Your server URL: in AnythingMCP, open MCP Servers → the server this connector is on, and copy its URL (https://cloud.anythingmcp.com/mcp/…). Use it wherever this guide shows YOUR_SERVER_ID.

ChatGPT adds MCP servers as plugins on its Plugins page, on the web and on paid plans. Developer mode is no longer needed. In a Business workspace only admins and owners can add one; in Enterprise and Edu an admin can grant the Create plugins with MCPs permission.

1. Go to chatgpt.com/plugins, click +, then Add custom MCP server.
2. Enter a name, e.g. Name: OCR.space.
3. Under Connection, enter the Server URL https://cloud.anythingmcp.com/mcp/YOUR_SERVER_ID and keep Authentication on OAuth.
4. Select I understand and want to continue, then Create as a plugin. When ChatGPT asks, sign in to AnythingMCP and allow access.
5. Install the new plugin from your personal plugins, then type @ in a chat and pick it.

Available tools

Example prompts

• "Read the receipt at https://example.com/receipts/2026-10-14.jpg and list every item with its price, plus the total."
• "Pull the invoice number, date, supplier and total amount out of the scanned invoice at https://example.com/scans/invoice-4711.pdf."
• "Turn the scanned contract at https://example.com/docs/lease-scan.pdf into a searchable PDF and give me the download link."
• "This photo of a delivery note is sideways: https://example.com/photos/delivery-note.jpg. Read it and tell me which items and quantities are on it."
• "Use the most accurate engine to read the handwritten meeting notes at https://example.com/notes/whiteboard.jpg."

FAQ

Does the free ChatGPT plan work with custom MCP servers? OpenAI documents custom MCP servers for paid plans only; the Free plan is not on its list.

Next steps

• Connector landing page
• Connect to Claude
• Connect to Cursor
• Connect to Gemini
• Connect to Meta Muse