Skip to content
Agents tracked: 492 Releases (7d): 545 Agent status: 1 with issuesSince last limit reset: Claude 9d · Codex 7d For AI agents: Markdown Updated Nov 20, 2026

AI agents ranked by real usageFind your agent.
Put it to work.

Which AI agent to use, what it costs and what changed. 492 agents ranked on live public data, refreshed daily: compare tools, understand the costs and start building with a setup that fits.

User reports and Answers Hit a bug?
See what others found.
Known issues, questions and fixes people report for AI agents and developer tools, from each project's GitHub. See user reports →

Explore AI agents

Full catalog →
All categories →
AgentVS
  1. Claude CodeTerminal · Anthropic 18.9M▲15% 7d 88
  2. OpenAI CodexTerminal · OpenAI 26.6M▲2% 7d 84
  3. OpenCodeTerminal · Anomaly 3.9M▲15% 7d 80
  4. n8nAutomation · n8n 107K▼28% 7d 79
  5. GitHub CopilotIDEs · GitHub — 75
  6. CursorIDEs · Anysphere — 73
  7. PiTerminal · Earendil Inc. 6.4M▲2% 7d 73
  8. OpenClawGeneral · OpenClaw Foundation 3.1M▲6% 7d 71
  9. ElevenLabs AgentsVoice · ElevenLabs 3M▲4% 7d 692/4
  10. DeepSeek HarnessGeneral · DeepSeek 658K▼22% 7d 67
  11. LovableApp builders · Lovable — 66
  12. ZedIDEs · Zed Industries — 64
  13. Hermes AgentGeneral · Nous Research 88.2K▲35% 7d 63
  14. DifyAutomation · LangGenius — 63
  15. Gemini CLITerminal · Google 495K▲17% 7d 62
  16. Browser HarnessBrowser · Browser Use 2.2M▼0% 7d 61
  17. Browser UseBrowser · Browser Use 2.2M▲11% 7d 60
  18. agent-browserBrowser · Vercel Labs 2.5M▲4% 7d 60
  19. Replit AgentApp builders · Replit — 60
  20. ClineTerminal · Cline Bot Inc. 84.3K▼9% 7d 58
  21. StagehandBrowser · Browserbase 2.9M▼13% 7d 57
  22. GitHub Copilot CLITerminal · GitHub 1.7M▼8% 7d 56
  23. MomenticTesting · Momentic 892K▲192% 7d 56
  24. DevinCloud · Cognition — 56
  25. Google AntigravityIDEs · Google — 55
  26. WarpIDEs · Warp — 55
  27. HerdrOrchestration · herdrdev — 55
  28. Oh My PiTerminal · Stencil Labs 313K▲30% 7d 54
  29. Retell AIVoice · Retell AI 809K▲13% 7d 542/4
  30. Open Code ReviewCode review · Alibaba 184K▲10% 7d 53
  31. PaperclipOrchestration · Paperclip 69.3K▲5% 7d 53
  32. Qwen CodeTerminal · Alibaba Cloud 155K▼11% 7d 53
  33. Kilo CodeTerminal · Kilo Code 60K▼3% 7d 53
  34. OrcaOrchestration · Stably — 53
  35. RasaSupport · Rasa 43K▼1% 7d 52
  36. LiveKit AgentsVoice · LiveKit 1.2M▼4% 7d 52
  37. VapiVoice · Vapi, Inc. 645K▲2% 7d 522/4
  38. ChatbaseSupport · Chatbase 865▲1960% 7d 512/4
  39. ManusGeneral · Butterfly Effect Pte. Ltd. — 51
  40. mini-SWE-agentTerminal · Princeton University & Stanford University 3.1M▼4% 7d 51
  41. e2eTesting · e2e 1.3M▲53% 7d 50
  42. KiroIDEs · Amazon Web Services — 50
  43. RufloOrchestration · rUv 69.4K▼28% 7d 50
  44. AnythingLLMAutomation · Mintplex Labs — 50
  45. AiderTerminal · Aider-AI 116K▲25% 7d 50
  46. Antigravity CLITerminal · Google — 49
  47. Amazon Q DeveloperIDEs · Amazon Web Services — 49
  48. GooseTerminal · Block — 49
  49. Bolt.newApp builders · StackBlitz — 47
  50. LangflowAutomation · Langflow 13.2K▼19% 7d 47

Gid Score (0–100) reflects adoption, community, attention and momentum, not output quality. Latest readings: Nov 19, 2026. How it works

Know what you’ll pay

AI agent pricing compared →

AI agents in action

Explore use cases →

How teams use AI agents: real results, guides, workflows and lessons learned.

Tools, APIs and skills for AI agents

Use AgentGid from your AI →

Extend your setup: everything an agent connects to or builds with, checked and ranked.

Your tools. Only the updates that matter.

Outage alerts, releases and price changes for the agents you use, plus a Monday digest.

Sign in to track tools →
Public data from npm, PyPI, GitHub, the VS Code Marketplace, vendor status pages and release notes · updated Nov 20, 2026 · What is AgentGid?

Browse by category

All categories →
AI coding agentsAgents that read a codebase, edit files, run commands and open pull requests, ranked on live adoption and act…87 trackedAI app buildersPrompt-to-app tools that generate a working application from a description, host it and let you iterate in ch…9 trackedAI agents for game developmentAgents that build and edit games inside the engine, MCP servers that let your coding agent drive Unity, Unrea…13 trackedAI design agentsAgents that turn a brief into editable designs and prototypes, apply your design system and revise on request.8 trackedAgent frameworks and SDKsLibraries for building your own agents: model calls, tools, memory, orchestration and multi-agent patterns.51 trackedAgent tooling for cost, context and memoryNot agents themselves, but the layers you add to one: cut token spend, give it memory, route and meter model …65 trackedAgent builders and workflow automationVisual and low-code platforms for wiring models, tools and triggers into agents and automated workflows.17 trackedGeneral-purpose AI agentsAgents that take an open-ended goal, such as research, a report or a booking, and carry it out across tools o…44 trackedBrowser and computer-use agentsAgents that operate a browser or a desktop the way a person does: reading pages, clicking, typing and filling…26 trackedAI research agentsAgents that do the reading for you: they search, screen sources, extract findings and write a report with cit…14 trackedAI data analysis agentsAgents that answer questions about your data by writing and running the analysis, and show their work.9 trackedAI agents for business teamsAgents that take on work in a business function, usually inside the systems that team already uses.76 trackedAI agents for DevOps and securityAgents that watch production and attack surfaces: they investigate alerts and incidents, find root causes and…17 trackedAgent orchestration and team workspacesWhen one agent is not enough: run Claude Code, Codex and others in parallel, hand work between them and keep …39 trackedAI video and media agentsAgents that turn a script, a topic or raw footage into finished video, and editors built so coding agents can…15 trackedAI trading and investing agentsAgents that read markets, research companies and run trading strategies. Most are research projects: test the…2 tracked

What this site tracks

AgentGid is a live scoreboard for AI agents: software that plans and carries out multi-step work with tools, such as editing a codebase, operating a browser or resolving a support ticket. Instead of opinions, it uses numbers anyone can check: how often an agent's packages are downloaded, how many people installed its editor extension, how fast its GitHub repository is gaining stars, how often it ships releases, and how much it is being talked about.

Each agent has its own page with history charts, pricing, the latest release notes and the closest alternatives. The full catalog ranks all 492 products, category pages rank agents built for the same job, and comparison pages put two of them side by side.

Building with agents? AgentGid also tracks the tools for building with AI agents that 6,254 real projects use, the MCP servers in the official registry, Agent Skills for Claude Code, Codex and Cursor, APIs for AI agents with their prices and free tiers, assets for apps and AI agents with checked licenses, AI model prices, usage limits and product recipes with the APIs, model and monthly cost of an AI product.

Frequently asked questions

What counts as an AI agent here?

Software that uses a language model to plan and carry out multi-step tasks with tools, rather than only answering in chat: coding agents that edit files and run commands, browser agents that click through websites, support agents that resolve tickets, and the frameworks developers use to build such systems.

How is the Gid Score calculated?

It is a 0 to 100 composite of four components: adoption (package downloads, extension installs, Homebrew installs, site traffic rank), community (GitHub stars), attention (Hacker News mentions and Wikipedia views) and momentum (30-day growth). Each signal is placed on a fixed logarithmic scale, and components without data are left out rather than counted as zero. The full formula is on the methodology page.

Are download counts the same as users?

No. Registries count every download, including CI pipelines, mirrors and automatic updates, so a tool that ships several releases a week is downloaded more often per user than one that ships monthly. Downloads are a good measure of activity and of direction over time, and a poor measure of how many people use a tool.

Why do some agents have no download or star numbers?

Closed products that are not distributed through public registries leave fewer public traces. For those we show what exists, such as site traffic rank, marketplace installs, status-page incidents and pricing, and mark the rest as unavailable instead of estimating it.

How often is the data updated?

AgentGid collects every source once a day. Download and page-view figures usually trail by one to two days because the registries publish them with a delay. Pricing and product facts are checked by hand and carry the date they were last verified.

Does AgentGid rank agents by quality?

Not directly. Usage and growth show what developers and companies are choosing, which is a strong signal but not the same as output quality. Where independent quality data exists, such as marketplace ratings or the share of an agent's pull requests that get merged, it is shown alongside the usage numbers.

Where do I find tools, MCP servers, skills and APIs for my AI agent?

In the catalogs next to the agent rankings: Tools for building with AI agents (the libraries real agent projects depend on), MCP servers from the official registry ranked by GitHub stars, Agent Skills packs for Claude Code, Codex and Cursor, APIs for AI agents with prices and free tiers, and assets for apps and AI agents with what each license allows.