Hugging Face Launches "Agent Hub" — A Marketplace for Sharing and Deploying AI Agents
Hugging Face has launched Agent Hub, a marketplace for discovering, sharing, and deploying AI agents. Similar to what the Model Hub did for ML models, Agent Hub provides one-click deployment, version control, and community ratings for autonomous AI agents.
Hugging Face has expanded its platform with Agent Hub, a new marketplace dedicated to AI agents. Just as the Hugging Face Model Hub became the go-to destination for sharing and discovering ML models, Agent Hub aims to do the same for autonomous AI agents.
Platform Features:
- •One-Click Deploy: Deploy any agent from the hub to Hugging Face Spaces with a single click, with automatic resource provisioning.
- •Agent Cards: Standardized documentation format describing an agent's capabilities, limitations, safety properties, required tools, and benchmark results.
- •Version Control: Full git-based versioning for agent code, prompts, and configurations.
- •Community Ratings: Users can rate agents on reliability, safety, and usefulness, with verified deployment badges.
- •Agent Leaderboard: Automated benchmarks ranking agents by task category (coding, research, data analysis, creative).
At launch, Agent Hub hosts over 2,000 agents built with frameworks including LangGraph, CrewAI, AutoGen, Smolagents, and custom implementations. Categories range from coding assistants to research agents to specialized domain agents for healthcare, legal, and finance.
"We believe in making AI agents accessible to everyone," said Clement Delangue, CEO of Hugging Face. "Agent Hub removes the barrier between building an agent and putting it in the hands of users."