slim.to integration docs
slim.to turns files and URLs into tracked short links — hosted viewer, open/viewer analytics, password protection, custom domains. These docs cover using slim.to from outside the web app: from AI assistants, scripts, and other tools.
The one-minute version
An AI assistant just wrote you a report, a pitch page, or a mockup. Instead of downloading the file and re-uploading it to slim.to, the assistant calls slim.to directly and hands you a tracked link:
"Here's your Q3 report: https://slim.to/4h7bqn — I'll be able to tell you when someone opens it."
There are two ways to wire that up:
| Approach | What it is | Best for |
|---|---|---|
| 1. MCP server | slim.to as native AI tools — a hosted connector at https://slim.to/mcp (OAuth, zero install) and a local stdio server (slimto-mcp) |
Hosted: claude.ai, ChatGPT connectors. Local: Claude Desktop / Claude Code, Gemini CLI, OpenAI Agents SDK |
| 2. Direct REST API | Personal API key + plain HTTPS calls (OpenAPI spec included) | Scripts, CI, custom GPT Actions, anything that can make an HTTP request |
Both authenticate with a personal API key — start there.
Guides
- API keys — create and manage keys (required for both approaches)
- MCP server — install and configure the
slimto-mcpserver - REST API quickstart — curl-level walkthrough of the API
Per-platform how-tos
Each guide covers both approaches for that platform:
- Claude — Claude Desktop, Claude Code, claude.ai
- OpenAI — ChatGPT, Agents SDK, custom GPT Actions
- Gemini — Gemini CLI, Google GenAI SDK
Limits
- Allowed file types:
pdf,png,jpg,jpeg,gif,html,zip,mp4 - Max file size: 100 MB (files over 30 MB use a two-step direct-to-storage upload)
- API key requests are rate-limited to 120/minute per key
- Free plan: 5 active links; Pro/Team lift the limits — see slim.to pricing