MCP Overview
The PrintHive MCP server lets AI assistants — Claude Code, Cursor, ChatGPT, and any MCP-compatible client — directly control and monitor your Bambu Lab print farm. It exposes 23 tools covering printer management, file operations, print control, hardware operations, filament handling, and telemetry.
Why MCP for Print Farms?
- Natural language control — "Pause the print on Printer #3" just works
- Batch operations — Ask your AI to check status across all printers at once
- Integrated workflow — Control printers without leaving your IDE or chat
- Full observability — Pull temperature telemetry, job history, and live status
Base URL
https://api.printhiv3d.com/v1/mcpThe server uses the Streamable HTTP transport (MCP spec 2025-03-26). All communication happens on the single endpoint above via POST, GET, and DELETE.
Authentication
The MCP server supports two authentication methods:
- API Key — Send the
x-api-keyheader with your organization API key. Generate one in Settings → API Keys. - Bearer Token — Send
Authorization: Bearer <token>with a valid Supabase JWT.
For AI assistant integrations, API keys are recommended since they don't expire and are scoped to your organization.
Tool Reference
All 23 tools available through the MCP server:
| Tool | Description |
|---|---|
| printers:list | List all printers in the organization |
| printer:live_status | Get real-time printer state (temps, progress, speeds, fans, AMS) |
| printer:ams | Get AMS filament details (types, colors, remaining) |
| printer:files | List files stored on the printer |
| printer:file_upload | Upload a model file from the library to a printer |
| printer:file_delete | Delete a file from a printer |
| printer:file_task_status | Check status of a file upload or delete task |
| jobs:list | List print jobs in the organization |
| jobs:get | Get detailed info about a specific print job |
| jobs:history | Get event history for a print job |
| telemetry:temperatures | Get temperature telemetry for a job |
| print:pause | Pause the current print |
| print:resume | Resume a paused print |
| print:stop | Cancel/stop the current print |
| print:set_speed | Set print speed (silent, normal, sport, ludicrous) |
| temperature:set | Set nozzle or bed temperature |
| gcode:send | Execute raw G-code commands |
| hardware:light | Toggle chamber or work light |
| hardware:fan | Set fan speed (part cooling, auxiliary, chamber) |
| hardware:home | Home all axes |
| hardware:auto_level | Run automatic bed leveling |
| filament:load | Load filament from an AMS tray |
| filament:unload | Unload filament from the extruder |
Client Setup Guides
How It Works
When an AI assistant connects to the Hive MCP server, it receives the list of available tools and their schemas. When you ask the AI to do something printer-related, it calls the appropriate tool with the right parameters. The MCP server authenticates the request, executes the action through the Hive API (and Pulse relay for live printer commands), and returns the result to the AI.
Live printer operations (status, control, hardware) are routed through the Pulse WebSocket relay to your local HiveLink instance, which communicates directly with your printers via MQTT. Database operations (jobs, history, telemetry) are read directly from Supabase.