Automate File Conversion: Convert.Online's API and MCP Server for AI Agents

You've Been Converting Files One at a Time. There's a Faster Way.

If you've ever needed to convert a hundred invoices to PDF, or resize a folder of product photos, you know the drill: open the site, upload one file, wait, download, repeat. It doesn't have to be that way anymore. Convert.Online now ships with a full REST API, a visual Job Builder that writes the code for you, and an MCP server that plugs conversion directly into Claude, ChatGPT, and Gemini.

Here's what each one actually does — and which one is right for you.

The API: Conversion as Infrastructure

The Convert.Online API is built around two simple ideas: jobs and tasks. A job is a single conversion request; tasks are the steps inside it — import, convert, export. That structure is what lets the same API handle a one-off PDF-to-Word conversion and a five-stage pipeline pulling a file from S3, converting it, and dropping the result on an SFTP server.

The quick-start flow is five calls:

1. Create the job — POST /v1/process/jobs
2. Upload your file — PUT the bytes, or import it from a URL or base64
3. Start the conversion — POST /complete-upload
4. Poll until finished — GET /v1/process/jobs/{id}, or subscribe to live job events over SSE instead of polling
5. Download the result

Beyond the basics, the API supports importing and exporting directly from S3, SFTP, and Google Cloud Storage, chaining multiple conversions into one job (PNG → PDF → JPG in a single request), and webhooks with signature verification so your backend gets notified the moment a job finishes — no polling loop required. There's also a sandbox mode for testing your integration without spending real quota.

Authentication is a single API key. No OAuth dance, no session cookies — just a key in a header and you're making requests.

API Job Builder: Skip the Documentation

Reading API docs is fine. Not reading them is faster. The API Job Builder is a visual tool — pick your source format, your target format, add the tasks you need — and it generates working, ready-to-run code in eight languages: cURL, Python, Node, PHP, Ruby, Java, Go, and more. Copy, paste, done.

It's the fastest way to find out whether the API fits your use case before you write a single line by hand.

MCP Server: Let Your AI Assistant Do the Converting

This is the one most file converters don't have. Convert.Online runs a full MCP (Model Context Protocol) server — the open standard that lets AI assistants like Claude, ChatGPT, and Gemini call real tools instead of just talking about them. Connect it once, and you can ask your assistant to convert a file mid-conversation. No tab-switching, no manual upload.

Connecting takes one of two paths:

Once connected, your assistant has five tools at its disposal:

• list_formats — see what conversions are supported
• list_options — check conversion-specific settings before running one
• create_upload — hand off a large local file for conversion
• convert — run the actual conversion
• get_job — check status and grab the result

Ask "convert this HEIC photo to JPG" or "turn this DOCX into a PDF" and it just happens — the assistant picks the right tool, runs it, and hands you the result.

API or MCP — Which One Do You Need?

Short version: if you're writing code that needs to convert files at scale — a backend job, a CI pipeline, a SaaS feature — use the API directly, or the Job Builder to get there faster. If you (or your team) already live inside an AI assistant and just want conversion to be one more thing it can do without a context switch, connect the MCP server and never think about it again.

Conclusion

Manual, one-file-at-a-time conversion was never going to scale — not for developers automating a pipeline, and not for anyone who'd rather just ask their assistant to handle it. Now it doesn't have to. Whether you're wiring up webhooks in production or asking Claude to batch-convert a folder of images, Convert.Online's API and MCP server turn file conversion from a task you do into infrastructure you don't think about.