Every project, by topic

All 11,207 projects the atlas merges, each with a page carrying its install command, its five platform verdicts and which of the 39 lists named it. Grouped by topic and alphabetical inside it — the star rankings are on the interactive index and the 14 topic pages.

Orchestrators & Multi-Agent · Coding Agents · Agent Skills · MCP Servers · Frameworks & SDKs · Harnesses & Runtime Infra · Context, Memory & RAG · Sandbox, Security & Governance · Observability & Evals · Plugins, Themes & Clients · Research & Data Agents · Creative, Voice & Media · Assistants & Domain Agents · Docs, Learning & Lists

Orchestrators & Multi-Agent 488

Running several agents at once, or one agent for a long time: swarms, loop runners, task queues, planners and the orchestration write-ups.

Orchestrators & Multi-Agent ranked by stars →

Coding Agents 487

The agent you point at a repository. Terminal, desktop and web, plus the parallel-agent front-ends built on top of them.

Coding Agents ranked by stars →

Agent Skills 1,453

Capability bundles an agent loads on demand — SKILL.md packages, subagents, slash commands, and the directories that collect them.

Agent Skills ranked by stars →

MCP Servers 5,180

The Model Context Protocol ecosystem: reference and third-party servers, inspectors, gateways, transports, and the sibling agent-to-agent protocols.

MCP Servers ranked by stars →

Frameworks & SDKs 504

Libraries you build an agent out of, and the tool-calling interfaces it reaches the world through.

Frameworks & SDKs ranked by stars →

Harnesses & Runtime Infra 686

The scaffolding around the model rather than the model or the agent: runtimes, providers, gateways, fine-tuning, and what a harness runs on in production.

Harnesses & Runtime Infra ranked by stars →

Context, Memory & RAG 156

Everything that decides what the model actually sees: retrieval, memory stores, compaction, and the chat-with-your-data apps built on them.

Context, Memory & RAG ranked by stars →

Sandbox, Security & Governance 184

Isolation, permissions, approval gates, guardrails, compliance and audit — plus the security agents that attack and defend.

Sandbox, Security & Governance ranked by stars →

Observability & Evals 453

Traces, benchmarks, graders, linting and CI feedback: how you find out whether the agent actually did the thing.

Observability & Evals ranked by stars →

Plugins, Themes & Clients 743

Attaches to an agent you already run — plugins, themes, status lines, alternative clients, notification bridges and generative UIs.

Plugins, Themes & Clients ranked by stars →

Research & Data Agents 42

Agents pointed at literature, the web and datasets: deep research, analysis, business intelligence and science.

Research & Data Agents ranked by stars →

Creative, Voice & Media 32

Image, audio, video, voice and prose agents, and the ones that play games.

Creative, Voice & Media ranked by stars →

Assistants & Domain Agents 653

One agent doing one job in one domain: support, sales, finance, health, travel, browser and desktop automation, personal assistants.

Assistants & Domain Agents ranked by stars →

Docs, Learning & Lists 146

Reading rather than running: guides, courses, pattern write-ups, practice postures, and the other awesome-lists this one was merged from.

Docs, Learning & Lists ranked by stars →