---
title: fastmcp3
emoji: 🔎
colorFrom: blue
colorTo: indigo
sdk: docker
app_port: 7860
short_description: fast-agent PR build serving the raw HF Hub query card.
---

# fastmcp3

This Space runs a raw-passthrough fast-agent MCP server for HF Hub queries, using a locally built fast-agent wheel from `~/source/fast-agent-pr/`.

The deployed card uses `tool_result_mode: passthrough`, so tool results are returned directly rather than rewritten by a second LLM pass.

## Auth

This Space is configured for Hugging Face OAuth/token passthrough:

- `FAST_AGENT_SERVE_OAUTH=hf`
- `FAST_AGENT_OAUTH_SCOPES=inference-api`
- `--instance-scope request`

These are configured as Space settings:

- Variables:
  - `FAST_AGENT_SERVE_OAUTH`
  - `FAST_AGENT_OAUTH_SCOPES`
  - `FAST_AGENT_OAUTH_RESOURCE_URL`
- Secret:
  - `HF_TOKEN` (dummy startup token)

Clients can either:
- send `Authorization: Bearer <HF_TOKEN>` directly, or
- use MCP OAuth discovery/auth flow

## Model

The deployed card uses:

- `hf.openai/gpt-oss-120b:sambanova`

## Main files

- `hf-hub-query.md` — raw MCP card
- `monty_api_tool_v2.py` — Hub query tool implementation
- `_monty_codegen_shared.md` — shared codegen instructions
- `wheels/` — optional local fast-agent wheel staging directory for one-off deploys

## Note on Monty

The Space now installs the released `pydantic-monty==0.0.8` package from PyPI, so the custom bundled Monty wheel is no longer required.