Everything GenAiHub does, in one place
GenAiHub is an open directory of AI agents, LLMs, MCP servers, workflows, frameworks, tools, and skills. It exists to get you from “what's out there?” to a decision you can defend — discover, evaluate, compare, choose. Here's how each stage works, plus submissions and your account.
From discovery to decision
- Discover— Find relevant AI resources.
Search by name, tagline, or tag from the home page or browse view — results update as you type. Narrow them with the browse facets (category, platform, pricing model, model type), which combine with AND, and reorder with the sort control above the grid: relevance, newest, most clicked, highest rated, most rated, or most liked. The top navigation also splits the catalog by type if you already know what you’re after.
Browse the catalog → - Evaluate— See capabilities, integrations and usage data.
Every listing page collects what you need to judge a resource in one scroll: description, capabilities, categories and tags, author, platform, and pricing model, plus performance benchmarks where the maintainer has published them. Community metrics, star ratings, and comments show how it has worked out for other people.
See published benchmarks → - Compare— Compare alternatives side by side.
Shortlist up to five listings and put them in a single table — pricing, platform, capabilities, and categories line up row by row, so the differences are the only thing left to read. Related listings on each detail page are a quick way to find candidates worth adding.
Open the comparison view → - Choose— Open docs, repository or deployment resources.
The Open/Install action on a live listing links straight to the source — docs, repository, or deployment resource, whatever the maintainer published. Not ready to commit? Save the listing or drop it into a collection and come back to it.
View saved agents →
The catalog at a glance
Every listing has a type, which drives its own navigation tab and icon:
- AgentsAutonomous AI agents that complete tasks end-to-end
- LLMsLarge language models callable via a hosted API or open weights
- MCP ServersModel Context Protocol servers exposing tools and resources
- WorkflowsMulti-step automation pipelines and orchestration templates
- FrameworksLibraries and SDKs for building AI-powered applications
- ToolsStandalone utilities and plugins that extend AI capabilities
- SkillsReusable SKILL.md capability packages that teach agents new tasks
Listing status
Every listing carries a status so you know whether it's safe to click through:
Available now. The Open/Install button links straight to the resource.
Temporarily unavailable pending review. Still visible, but the install action is disabled.
Reported or detected as unreachable. Shown with a broken-link indicator instead of an install action.
Submitting, saving, and getting involved
Anyone can submit an agent, LLM, MCP server, or tool from the Submit page. New submissions start in review and go live once approved.
Submit an agent →Sign in and click Save on any listing to bookmark it. Your saved list is always one click away from the account menu.
View saved agents →Group listings into named collections and set them private, unlisted, or public so they can be shared with a link.
View your collections →Leave a 1–5 star rating or a quick like on any listing. Aggregate scores feed the "Highest rated" and "Most liked" sort options.
Browse and rate agents →Spotted a broken link, spam, or inaccurate listing? Report it straight from the listing page and track the outcome afterwards.
Report a listing →Moderation
Submissions and reports both go through review before they affect what's visible on the site. Approved submissions go live; reports are triaged and can result in a listing being corrected, suspended, or removed.