How do I get a Hugging Face Access Token?

Log in to [**Hugging Face**](https://huggingface.co), go to **Settings > Access Tokens**, click **New token**, give it a name and select scopes (read is sufficient for browsing, write if you need to create repos). Copy the token immediately — it starts with `hf_`.

Can I search models by task type (e.g. text-generation)?

Yes! Use `list_models` with a search query. While the search endpoint doesn't directly filter by pipeline_tag, you can search by task name (e.g. search='text-generation') and then use `get_model` or `get_model_tags` to verify the pipeline_tag of specific models.

Can I see what files are in a model repository?

Yes! Use `list_model_files` with the model ID (e.g. 'google-bert/bert-base-uncased') to see the complete file tree including model weights (.safetensors, .bin), config files, tokenizer files and README. Optionally set a path to browse a specific subdirectory like 'onnx' or 'pytorch'.

Can I create discussions on model pages?

Yes! Use `create_discussion` with the repo type ('model', 'dataset' or 'space'), the repo ID and a title. This creates a new discussion thread on the repository. You can use `list_model_discussions` first to check existing threads before creating a new one.

Hugging Face MCP Server for Pydantic AI 13 tools — connect in under 2 minutes

Name: Hugging Face
Availability: InStock
Author: Vinkius

ai.pydantic.dev/mcp

Built by Vinkius GDPR 13 Tools SDK

Pydantic AI brings type-safe agent development to Python with first-class MCP support. Connect Hugging Face through the Vinkius and every tool is automatically validated against Pydantic schemas — catch errors at build time, not in production.

Get MCP Server for AI Agents

ASK AI ABOUT THIS MCP SERVER

Open in ChatGPT Open in Claude Open in Perplexity

Vinkius supports streamable HTTP and SSE.

python

import asyncio
from pydantic_ai import Agent
from pydantic_ai.mcp import MCPServerHTTP

async def main():
    # Your Vinkius token — get it at cloud.vinkius.com
    server = MCPServerHTTP(url="https://edge.vinkius.com/[YOUR_TOKEN_HERE]/mcp")

    agent = Agent(
        model="openai:gpt-4o",
        mcp_servers=[server],
        system_prompt=(
            "You are an assistant with access to Hugging Face "
            "(13 tools)."
        ),
    )

    result = await agent.run(
        "What tools are available in Hugging Face?"
    )
    print(result.data)

asyncio.run(main())

Fully ManagedVinkius Servers

60%Token savings

High SecurityEnterprise-grade

IAMAccess control

EU AI ActCompliant

DLPData protection

V8 IsolateSandboxed

Ed25519Audit chain

<40msKill switch

Stream every event to Splunk, Datadog, or your own webhook in real-time

* Every MCP server runs on Vinkius-managed infrastructure inside AWS - a purpose-built runtime with per-request V8 isolates, Ed25519 signed audit chains, and sub-40ms cold starts optimized for native MCP execution. See our infrastructure

About Hugging Face MCP Server

Connect your Hugging Face account to any AI agent and explore the world's largest AI model hub through natural conversation.

Pydantic AI validates every Hugging Face tool response against typed schemas, catching data inconsistencies at build time. Connect 13 tools through the Vinkius and switch between OpenAI, Anthropic, or Gemini without changing your integration code — full type safety, structured output guarantees, and dependency injection for testable agents.

What you can do

Model Discovery — Search and browse thousands of models by name, task type, framework and author
Model Inspection — View model metadata including pipeline task, tags, download counts, likes and file structure
Dataset Exploration — Find and inspect datasets with their descriptions, sizes and file trees
Spaces Gallery — Browse ML demo apps (Gradio, Streamlit, Docker) and check their runtime status
Collections — View curated collections of models, datasets and spaces organized by topic
Community Discussions — Read model discussion threads for bug reports, feature requests and usage tips
File Tree Browsing — List repository files (model weights, configs, tokenizers) without downloading

The Hugging Face MCP Server exposes 13 tools through the Vinkius. Connect it to Pydantic AI in under two minutes — no API keys to rotate, no infrastructure to provision, no vendor lock-in. Your configuration, your data, your control.

How to Connect Hugging Face to Pydantic AI via MCP

Follow these steps to integrate the Hugging Face MCP Server with Pydantic AI.

Install Pydantic AI

Run pip install pydantic-ai

Replace the token

Replace [YOUR_TOKEN_HERE] with your Vinkius token

Run the agent

Save to agent.py and run: python agent.py

Explore tools

The agent discovers 13 tools from Hugging Face with type-safe schemas

Why Use Pydantic AI with the Hugging Face MCP Server

Pydantic AI provides unique advantages when paired with Hugging Face through the Model Context Protocol.

Full type safety: every MCP tool response is validated against Pydantic models, catching data inconsistencies before they reach your application

Model-agnostic architecture — switch between OpenAI, Anthropic, or Gemini without changing your Hugging Face integration code

Structured output guarantee: Pydantic AI ensures tool results conform to defined schemas, eliminating runtime type errors

Dependency injection system cleanly separates your Hugging Face connection logic from agent behavior for testable, maintainable code

Hugging Face + Pydantic AI Use Cases

Practical scenarios where Pydantic AI combined with the Hugging Face MCP Server delivers measurable value.

Type-safe data pipelines: query Hugging Face with guaranteed response schemas, feeding validated data into downstream processing

API orchestration: chain multiple Hugging Face tool calls with Pydantic validation at each step to ensure data integrity end-to-end

Production monitoring: build validated alert agents that query Hugging Face and output structured, schema-compliant notifications

Testing and QA: use Pydantic AI's dependency injection to mock Hugging Face responses and write comprehensive agent tests

Hugging Face MCP Tools for Pydantic AI (13)

These 13 tools become available when you connect Hugging Face to Pydantic AI via MCP:

create_discussion

Requires the repo type (model, dataset or space), the repo ID in "author/name" format and the discussion title. Returns the created discussion with its ID, title and URL. Create a new discussion on a Hugging Face repo

get_collection

Provide the collection slug. Get details for a specific Hugging Face collection

get_model

Provide the model ID in "author/name" format (e.g. "google-bert/bert-base-uncased"). Get details for a specific Hugging Face model

get_model_tags

Tags include framework (pytorch, tensorflow), license, dataset, language and task-specific labels. The pipeline_tag indicates the model's primary task (e.g. "text-generation", "image-classification", "translation"). Get tags and pipeline info for a Hugging Face model

get_space

Provide the space ID in "author/name" format. Get details for a specific Hugging Face Space

get_user

Returns user name, avatar, organizations, auth type, plan and access tokens metadata. Use this to verify your token is working correctly. Get the authenticated Hugging Face user

list_collections

Optionally filter by author and limit. Returns collection slug, title, description, author, item count and likes count. List collections on Hugging Face Hub

list_dataset_files

Returns filenames (e.g. "train.parquet", "test.parquet", "data/", "README.md"). Optionally set a subdirectory path. Useful for understanding dataset structure before downloading. List files in a Hugging Face dataset repository

list_datasets

Optionally filter by search term, author and limit. Returns dataset ID, author, description, download count, likes count and creation date. List datasets on Hugging Face Hub

list_model_discussions

Returns discussion title, author, creation date, number of comments and whether it is resolved. Use this to review community feedback, bug reports and feature requests for a model. List discussions for a Hugging Face model

list_model_files

Returns filenames, file sizes and paths (e.g. "model.safetensors", "tokenizer.json", "config.json", "README.md"). Optionally set a subdirectory path to list files within a specific folder. Useful for inspecting model artifacts and understanding the repository structure. List files in a Hugging Face model repository

list_models

Optionally filter by search term (free-text across model cards), author (organization or username) and limit the number of results. Returns model ID, author, pipeline task tag, download count, likes count and creation date. List models on Hugging Face Hub

list_spaces

Optionally filter by search term, author and limit. Returns space ID, title, author, SDK (Gradio, Streamlit, Docker), likes count and creation date. List Spaces on Hugging Face Hub

Example Prompts for Hugging Face in Pydantic AI

Ready-to-use prompts you can give your Pydantic AI agent to start working with Hugging Face immediately.

"Find popular text generation models with over 1000 likes."

"Show me what files are in the bert-base-uncased model."

"What discussions are happening on the Llama-3 model page?"

Troubleshooting Hugging Face MCP Server with Pydantic AI

Common issues when connecting Hugging Face to Pydantic AI through the Vinkius, and how to resolve them.

MCPServerHTTP not found

Update: pip install --upgrade pydantic-ai

Hugging Face + Pydantic AI FAQ

Common questions about integrating Hugging Face MCP Server with Pydantic AI.

How does Pydantic AI discover MCP tools?

Create an MCPServerHTTP instance with the server URL. Pydantic AI connects, discovers all tools, and generates typed Python interfaces automatically.

Does Pydantic AI validate MCP tool responses?

Yes. When you define result types as Pydantic models, every tool response is validated against the schema. Invalid data raises a clear error instead of silently corrupting your pipeline.

Can I switch LLM providers without changing MCP code?

Absolutely. Pydantic AI abstracts the model layer — your Hugging Face MCP integration works identically with OpenAI, Anthropic, Google, or any supported provider.