TabPFN Skill

Product summary

TabPFN is a pre-trained transformer foundation model for tabular data that delivers high-quality predictions in seconds without dataset-specific training. It learns from billions of synthetic datasets and applies in-context learning to your data via a single forward pass. Use it via the Python client (tabpfn-client for API, tabpfn for local), REST API, or Model Context Protocol (MCP) for AI assistants.

Key entry points:

• Python API client: tabpfn-client (managed compute, no GPU needed) or tabpfn (local inference, GPU recommended)
• REST API: /tabpfn/* JSON endpoints at https://api.priorlabs.ai
• MCP server: https://api.priorlabs.ai/mcp/server for Claude, ChatGPT, Cursor, n8n
• Primary docs: https://docs.priorlabs.ai

Current model families: TabPFN-3.5 (base, Plus, Thinking, Fast), TabPFN-3 (base, Plus, Thinking), TabPFN-2.6, TabPFN-2.5, TabPFNv2.

When to use