Resize Baba for ChatGPT, Claude, Gemini and other AI assistants
AI assistants cannot resize or compress a file on their own. Resize Baba gives them the tools: connect the MCP server at https://resizebaba.com/mcp, or import the OpenAPI file at https://resizebaba.com/openapi.json. The assistant can then resize a photo to an exact KB, convert formats, build a PDF or make a passport sheet, and hand you a download link.
Open the OpenAPI file →Free · no signup · your files never leave your device
What an assistant can do
- list_size_presets: find the exact size for SSC, UPSC, IBPS, NEET, JEE, SBI, GATE, Amazon and more.
- resize_image: a photo or signature to an exact KB window and pixel size, verified, as a JPG or WebP.
- convert_image: JPG, PNG or WebP, with rotate and flip; camera and location data removed.
- images_to_pdf and merge_pdfs: build one PDF from photos or several PDFs.
- passport_photo_sheet: a printable 4x6 or A4 sheet of passport photos (needs a Pro key).
Connect Claude
- Open Settings, then Connectors, then Add custom connector.
- Name it Resize Baba and enter the URL https://resizebaba.com/mcp.
- Ask Claude: 'Resize this photo for the SSC form' and attach or link the picture.
Connect ChatGPT
- Custom GPT: Create, then Configure, then Actions, then Import from URL, and enter https://resizebaba.com/openapi.json.
- If your plan supports custom connectors, you can instead add the MCP URL https://resizebaba.com/mcp.
- Ask: 'Reduce this image to under 50 KB' and give a link or upload the file.
Gemini and other assistants
Any assistant or app that supports MCP can use https://resizebaba.com/mcp. Any that supports OpenAPI or function calling can use the file at https://resizebaba.com/openapi.json: pass the tool definitions to the model and call our endpoint when it asks. Each tool is one POST under /api/agent/v1/.
Call it directly
Send JSON to a tool. You can give the picture as a public link (image_url) or as base64 (image_base64).
curl -X POST https://resizebaba.com/api/agent/v1/resize_image \
-H "content-type: application/json" \
-d '{"image_url": "https://example.com/photo.jpg", "preset_id": "sscp"}'What comes back
Every result carries a download link. An assistant should show it to the user; the file is removed after 15 minutes.
{
"ok": true,
"download_url": "https://resizebaba.com/api/agent/v1/files/...",
"size_kb": 31.4, "width": 350, "height": 450,
"met_target": true, "verified": true,
"expires_in_minutes": 15
}Limits and Pro keys
- Free: about 20 operations a day per person, 6 a minute, files up to 5 MB, exam and general presets.
- Pro key: no daily cap, files up to 15 MB, marketplace presets such as Amazon and Flipkart, and passport photo sheets.
- Create a key in your account under Integrations, then send it as 'Authorization: Bearer rsz_...' (in a Custom GPT, choose API key, Bearer).
Privacy
The website tools never upload your files. The agent API is different, because an assistant cannot reach files on your device: the file is processed on our server, kept for 15 minutes behind a random link, then deleted. We do not read these files or use them to train anything. See the privacy policy.
Questions people ask
Why can't ChatGPT or Gemini resize an image to 50 KB by itself?
Chat assistants work with text and cannot run image tools on your files unless they are connected to one. Resize Baba is that tool.
Is it free for an assistant to use?
Yes, within a small daily allowance. A Pro key removes the cap.
Which file types work?
JPG, PNG, WebP, GIF, AVIF and HEIC images, and PDFs for merging. Results are JPG, WebP, PNG or PDF.
How do I get the file the assistant made?
The assistant gives you a download link. Open it within 15 minutes.