What AgentRank is.
AgentRank (agentrankhq.com) is a benchmark of what AI coding agents actually install, build with, and keep. We run the real agent CLIs against real repos, hundreds of sessions per category, and measure the four things that decide a developer tool's standing with agents: how often it gets named, installed, compiled, and kept.
What we are not.
The name gets shared with other projects, so to be precise: AgentRank is not a directory of AI agents, not a leaderboard comparing coding assistants, and not an MCP-server ranking. We do not rank agents. We measure how agents treat developer tools, and we help the vendors behind those tools track and improve their standing.
Why it exists.
AI coding agents already decide a growing share of tool adoption. When a developer says "add auth" and the agent picks a provider, installs it, and ships it, that selection never touches a search results page. Existing visibility tools measure what AI assistants say. AgentRank measures what agents do: the install, the build result, and whether the integration survives the next refactor. Actions, not mentions.
Facts.
- Founded
- 2026
- Website
- agentrankhq.com
- What it measures
- Agent Reach, First-Install Rate, Build Success, Refactor Retention
- Agents covered
- Claude Code and Codex, split by model
- Method
- Deterministic detection, pre-registered classifiers, n >= 30 with 95% Wilson intervals
- Contact
- [email protected]
The rules we hold ourselves to.
- Deterministic detection. Outcomes are read from the repo, never inferred, and never graded by another LLM.
- Public, versioned methodology. The method is open so the numbers can be checked.
- No sponsored placement, ever. Standing on AgentRank cannot be bought.
- Synthetic benchmarks, always labeled. Controlled runs are reported as directional, with sample size and a 95% Wilson interval on every share.
- Aggregates are public; a vendor's full numbers are private to that vendor.