AGENT ACTIVITY

Which AI agents developers actually use

Daily rankings of real API requests to AI agents and apps, plus a 13-week trend in tool calls. No estimates, no surveys — just routed traffic.

What this measures — read before citing

These are real requests and tool calls routed through OpenRouter. They are NOT unique users, NOT global market share, and exclude direct enterprise contracts. OpenRouter traffic skews toward individual developers and open-weight models.

Tool calls, by week

Tool calls are the clearest public signal of agentic usage — an agent has to call tools to do work.

+6.2% vs previous complete weeks
2026-08-10: 732,508,794 tool calls2026-08-17: 886,652,436 tool calls2026-08-24: 1,008,787,976 tool calls2026-08-31: 1,172,597,506 tool calls2026-09-07: 1,540,941,949 tool calls2026-09-14: 1,241,431,941 tool calls2026-09-21: 1,291,802,006 tool calls2026-09-28: 1,575,999,132 tool calls2026-10-05: 1,576,513,134 tool calls2026-10-12: 1,392,673,847 tool calls2026-10-19: 1,555,688,545 tool calls2026-10-26: 1,651,657,777 tool calls2026-11-02: 1,206,370,476 tool calls08-1009-2111-02

2026-08-10 → 2026-11-02 · latest week may be incomplete

Agent / App leaderboard

Ranked by total API requests. Switch the window to compare short-term spikes with sustained usage.

01
Portkey AI ↗Programming

Control panel for AI apps

16.8M

requests

02
Hermes Agent ↗Personal agentCLI agent

Hermes Agent is an open-source, self-improving AI agent by Nous Research that runs persistently with memory across sessions, and builds reusable skills from experience. It comes with 40+ built-in tools, including web sea

16.0M

requests

03

Claude Code is Anthropic's agentic coding tool that reads your entire codebase, plans and executes changes across files, runs tests, and iterates on failures, all from natural language prompts.

15.5M

requests

04
omp ↗CLI agent

11.3M

requests

05
Codex ↗CLI agent

A coding agent that helps you build and ship with AI

7.8M

requests

06
pi ↗CLI agent

There are many coding agents, but this one is yours.

6.6M

requests

07
Kilo Code ↗CLI agentide-extension

Kilo Code is an open-source AI coding agent that works across VS Code, JetBrains, and CLI to help developers ship code faster with agentic workflows.

6.3M

requests

08

5.1M

requests

09
OpenClaw ↗Personal agentCLI agent

OpenClaw is an open-source AI agent that connects to your messaging apps and takes real actions on your behalf, from running commands and browsing the web to managing files and sending emails.

4.1M

requests

10
Strix ↗CLI agent

Open-source AI hackers that find and validate vulnerabilities in your code, APIs, and infrastructure, and open fix PRs.

2.1M

requests

11

2.1M

requests

12
OpenHands ↗CLI agent

AI coding agent that can run commands, browse the web, call APIs

1.6M

requests

13
Cline ↗ide-extensionCLI agent

Cline is an open-source AI coding agent that lives inside your IDE, autonomously exploring your codebase, editing files, running terminal commands, and using browser automation.

1.4M

requests

14

1.3M

requests

15

AI adventures. Travel with your favorite characters

1.3M

requests

16

1.1M

requests

17

Powerful coding models, funded by ads

821.6K

requests

18
Proto Agent ↗Personal agentgeneral-chat

Proto is the personal agent for your business.

736.6K

requests

19

492.1K

requests

20
CodeGPT ↗ide-extensionCLI agent

366.5K

requests

More signals, kept separate

A single source can be skewed. These are independent signals from OpenRouter, npm/PyPI, GitHub and Artificial Analysis — shown separately so each can be checked on its own.

4/4 sources live

HOW TO READ: Independent signals, shown separately. No weighted composite score. NOT unique users, NOT global market share. Each panel measures a different thing.

Model-level API requests

OpenRouter · 2026-11-06

Real requests routed through OpenRouter, aggregated by model. Complements the Agent / App ranking above.

deepseek/deepseek-v4.1-flash-202610171.39B
typesafe/jev-1.13-202609171.05B
deepseek/deepseek-v4-flash-20260731652.7M
z-ai/glm-5.3-flash-20260826559.3M
openai/gpt-6-luna-20260922534.4M
typesafe/jev-1.14-20261014492.4M
anthropic/claude-haiku-5.5-20261007199.1M
perplexity/pplx-decider-v1.2-20261016182.4M
openai/gpt-5.6-luna-20260709163.3M
openai/text-embedding-3-small151.7M

Developer integration (SDK downloads)

npm / PyPI · 2026-11-06

Weekly downloads of major AI SDKs on npm and PyPI, with a 90-day npm trend below.

npm

@anthropic-ai/sdk65.7M
openai56.2M
ai44.3M
@google/genai31.6M
langchain4.1M

PyPI

openai91.2M
anthropic50.1M
langchain36.2M
google-generativeai2.5M
llama-index410.0K
2026-08-082026-11-06
openai@anthropic-ai/sdk@google/genai

Each line is normalized to its own maximum — compare shapes, not absolute heights. The weekly sawtooth is real: weekday downloads are far higher than weekends.

Open-source attention (GitHub stars)

GitHub · 2026-11-06

Stars of major AI SDK / framework repositories.

ollama/ollama186.1K
huggingface/transformers168.6K
langchain-ai/langchain149.1K
openai/openai-python32.1K
vercel/ai27.8K
anthropics/anthropic-sdk-python4.0K

Capability reference (NOT usage)

Artificial Analysis · 2026-11-06

Artificial Analysis Intelligence Index. Included as context only — it measures model quality, not how much a model is used.

Gemini 4 Carbon60
Claude Opus 5.559.1
GPT-6 Astra56.9
GPT-6.1 Sol54.6
Claude Sonnet 5.554.5
GPT-6.1 Terra51.3
Muse Wonder Full50.1
Claude Fable 5.149.2

Cross-validated vendors

Vendors appearing in more than one independent signal. Presence only — not a score or a ranking.

anthropic · 4 signalsopenai · 4 signalsgoogle · 3 signalsalibaba · 2 signalsdeepseek · 2 signalslangchain · 2 signalsmeta · 2 signals

LIMITATIONS

  • ·OpenRouter figures are real API requests routed through OpenRouter, excluding direct enterprise contracts.
  • ·npm/PyPI downloads include CI retries, mirrors and automated installs; they reflect SDK integration activity, not API usage.
  • ·GitHub stars measure attention/interest, not usage.
  • ·Artificial Analysis Intelligence Index measures model capability, not usage. Shown only as context.
  • ·Data sources are independent; do not add their numbers together.

DATA SOURCE & LIMITATIONS

  • ·OpenRouter skews toward individual developers and open-weight models; excludes direct enterprise contracts.
  • ·SDK/CI retries and automated agents are counted as requests.
  • ·Ranking is recomputed from total_requests; upstream rank field may have gaps.

OpenRouter public rankings endpoints (/api/frontend/v1/rankings/apps, .../tools) · verified 2026-11-06