AI Agent Landscape

A curated, opinionated map of the AI agent ecosystem.
Find the right tool for coding, browser automation, research, workflows, personal assistance, and agent development.

Start Here • Featured Picks • Comparison Snapshot • Categories • Use-Case Guides • How This Repo Thinks • Contributing

Why this repo exists

The AI agent space is moving insanely fast.

That sounds exciting until you actually try to choose a tool.

Then it becomes:

• endless launch tweets
• shallow "top 10" lists
• stale comparisons
• weak distinctions between products, frameworks, assistants, and experiments

AI Agent Landscape is meant to be the practical map.

Not the biggest list. Not the emptiest directory. Not a dressed-up SEO farm.

The goal is simple: help people find the right agent faster.

The repo is already tracking 100+ tools across coding, browser automation, research, workflows, app building, agent memory, protocols, assistants, and frameworks.

⚡ Start Here

If you only spend 30 seconds here, use this flow:

You want an agent to write or fix code

Start with Claude Code, Codex CLI, Cursor, Aider, OpenHands, or Cline.

You want an agent to act on websites

Start with Browser Use, Skyvern, or Stagehand.

You want deep web research and synthesis

Start with Perplexity, OpenAI Deep Research, Gemini Deep Research, or Genspark.

You want to build your own agent systems

Start with LangGraph, AutoGen, PydanticAI, smolagents, Semantic Kernel, or Mastra.

You want workflow automation with agent behavior

Start with CrewAI, n8n, Flowise, or Dify.

You want a personal or operator-style assistant

Start with OpenClaw, Open Interpreter, Khoj, or Letta.

You want to build apps with AI (not code them)

Start with Bolt.new, v0.dev, Lovable, or Replit Agent.

You want your agents to have memory

Start with Mem0, Zep, or Letta.

🔥 Featured Picks

These are not "the winners." They are the tools that currently stand out for clear reasons.

📊 Comparison Snapshot

🗂 Categories

Coding Agents

Tools focused on writing, editing, refactoring, reviewing, or debugging code.

Examples: Claude Code, Codex CLI, Cursor, GitHub Copilot, Aider, OpenHands, Cline, Continue, Goose, Gemini CLI, Devin.

Browser Agents

Tools that act on websites and web apps by browsing, clicking, extracting, and completing tasks.

Examples: Browser Use, Skyvern, Stagehand.

Research Agents

Tools optimized for search, synthesis, citations, and broad investigation.

Examples: Perplexity, OpenAI Deep Research, Gemini Deep Research, Genspark, You.com research workflows.

Workflow Agents

Tools that connect apps, APIs, documents, and actions into agentic business or operator flows.

Examples: CrewAI, n8n, Flowise, Dify.

Personal Assistants

Agents that help across messages, tasks, files, notes, browser actions, and devices.

Examples: OpenClaw, Open Interpreter, Khoj, Letta.

Agent Frameworks

Building blocks for developers who want to create custom agents and orchestrations.

Examples: LangGraph, AutoGen, Semantic Kernel, PydanticAI, Mastra, smolagents, LlamaIndex.

App Builder Agents

Tools that generate full applications or production-ready UI components from natural language prompts.

Examples: Bolt.new, v0.dev, Lovable, Replit Agent.

Agent Memory & Context

Tools and layers that provide persistent memory, long-term context, and knowledge management for AI agents.

Examples: Mem0, Zep, Letta.

Agent Protocols & Standards

Open standards and protocols that enable agent-to-tool and agent-to-agent communication.

Examples: Model Context Protocol (MCP), Google A2A.

📚 Use-case guides

• Best by Use Case
• Methodology
• Maintainer Notes

🧭 How to use this repo

This repo is for four kinds of people:

1. Developers choosing tools

You want the shortest path to the right agent for a real job.

2. Founders and operators

You want to understand which tools are real, which are hype, and which are worth integrating.

3. Builders comparing ecosystems

You want to see how products, frameworks, and open-source projects relate to each other.

4. Contributors and curators

You want to improve the map with better entries, better sources, and sharper distinctions.

🧠 How this repo thinks

This repo is curated and opinionated.

That means:

• we care more about usefulness than raw count
• we do not want a giant junk drawer of AI links
• we prefer short, honest notes over marketing language
• we want entries to answer "what is this actually good at?"

Read the full approach in docs/methodology.md.

📁 Repo structure

.
├── README.md
├── LICENSE
├── CONTRIBUTING.md
├── SECURITY.md
├── SUPPORT.md
├── assets/
│   ├── hero.svg
│   └── logo.svg
├── data/
│   ├── agents.csv
│   └── categories.md
├── docs/
│   ├── best-by-use-case.md
│   ├── maintainer-notes.md
│   └── methodology.md
└── .github/
    └── ISSUE_TEMPLATE/

🛣 Roadmap

• [x] Ship a credible public repo structure
• [x] Add visual identity assets
• [x] Publish an initial comparison snapshot
• [x] Expand the dataset to 75+ strong entries → now 100+
• [ ] Add sharper pricing / hosting / licensing metadata
• [x] Add category-specific shortlists and buying guides
• [ ] Add generated website / docs experience
• [ ] Add changelog and update cadence
• [x] Add "best by use case" pages
• [ ] Add Agent Protocols section
• [ ] Track MCP adoption across tools
• [ ] Monthly update cadence

🔐 Trust and safety

This repo is intentionally low-risk to browse and contribute to.

Right now it contains:

• Markdown docs
• SVG assets
• CSV data
• GitHub templates

It does not contain:

• install scripts
• runtime services
• package hooks
• GitHub Actions workflows

More detail: SECURITY.md

🤝 Contributing

Contributions are welcome, especially if you can help with:

• missing tools worth tracking
• stale pricing or licensing info
• category cleanup
• better descriptions
• clearer sourcing
• more useful comparison criteria

Before opening a PR, please read CONTRIBUTING.md.

⭐ Star this repo

If this repo saves you time, helps you choose faster, or gives you a better mental model of the agent ecosystem, give it a star.

License

MIT