How does AgentRank work?
AgentRank weighs verified adoption evidence, public product evidence, projects, saves, comparisons, usage signals, retention, and trust.
Verified adoption graph rankings for AI tools, agents, MCP servers, and developer platforms. Rankings combine activation signals, trust, protocol support, benchmarks, and commercial readiness.
Claim your profile to verify the AgentRank badge, add usage evidence, and see the developers behind each adoption signal.
Claim your toolActually adoptedAgentRank weighs verified adoption evidence, public product evidence, projects, saves, comparisons, usage signals, retention, and trust.
No. Sponsored placement can be labeled elsewhere, but AgentRank is not pay-to-play.
Stars and traffic show attention. Adoption proof shows which products developers put into real workflows.
Freshness
Live graph
Confidence
Usage, retention, trust, and protocol evidence are scored separately.
Privacy
Cohorts below N=7 stay private; those rows rank on public signals until they clear the threshold.
Methodology
Ranks link back to the published measurement model.
30-day adoption change; ★ shows GitHub star change where adoption is unavailable.
| Rank | Tool | AgentRank | Growth · 30d | Trend | Protocols | Trust |
|---|---|---|---|---|---|---|
| #1 | HKUDS/nanobot AI AgentsPublic signals | 52 | — | ★ 50k | API, SDK | observed · 49 |
Rankings are designed to become the public scoreboard for developer adoption. They weigh activation signals, public product evidence, projects, saves, comparisons, and retention over review volume or sponsored placement. Cohorts below N=7 are explicitly suppressed.
Recorded activation signals
Trust and telemetry proof
Protocol metadata
Benchmark performance
Commercial readiness
Non-pay-to-play policy
Public rankings are free and cannot be bought. Claiming and instrumenting your product can improve the quality of the public proof, then enterprise plans unlock the private account, cohort, and benchmark intelligence behind it.
Start with a free company workspace. Paid enterprise access is enabled manually through sales after agreement—there is no self-checkout.
| #2 | Playwright MCP AI AgentsPublic signals | 52 | +1.4k ★ | ★ 39k | API, SDK | observed · 49 |
| #3 | Supabase MCP Server Agent FrameworksPublic signals | 52 | +59 ★ | ★ 3k | API, SDK | observed · 49 |
| #4 | IBM MCP Context Forge Agent FrameworksPublic signals | 52 | +194 ★ | ★ 4.8k | API, SDK | observed · 49 |
| #5 | FastAPI-MCP Agent FrameworksPublic signals | 52 | +9 ★ | ★ 12k | API, SDK | observed · 49 |
| #6 | Apify MCP Server Agent FrameworksPublic signals | 52 | +5.1k ★ | ★ 15k | API, SDK | observed · 49 |
| #7 | mcp-agent Agent FrameworksPublic signals | 52 | — | ★ 8.6k | API, SDK | observed · 49 |
| #8 | MCP Toolbox for Databases Backend & DatabasePublic signals | 52 | +193 ★ | ★ 17k | API, SDK | observed · 49 |
| #9 | E2B Cloud DeployPublic signals | 51 | +522 ★ | ★ 15k | API, SDK | observed · 47 |
| #10 | Cog Cloud DeployPublic signals | 51 | +5 ★ | ★ 9.5k | API, SDK | observed · 47 |
| #11 | Browserless Cloud DeployPublic signals | 51 | +132 ★ | ★ 14k | API, SDK | observed · 47 |
| #12 | BentoML Cloud DeployPublic signals | 51 | +45 ★ | ★ 8.9k | API, SDK | observed · 47 |