# AI Agent Infrastructure Directory — agentinfra.dev

> A bilingual (English / 中文), fact-checked directory of 80 AI Agent infrastructure tools across 8 categories: runtime & orchestration, execution sandbox, observability, identity & permission, payment, MCP tooling, and memory. Every entry carries a self-contained, citable judgment plus a verification date.

## Core Pages
- https://agent.c8.fit/categories/personal-agent/: Personal Agents — 6 tools
- https://agent.c8.fit/categories/runtime/: Runtime & Orchestration — 19 tools
- https://agent.c8.fit/categories/sandbox/: Execution Sandbox — 15 tools
- https://agent.c8.fit/categories/observability/: Observability — 10 tools
- https://agent.c8.fit/categories/identity/: Identity & Permission — 8 tools
- https://agent.c8.fit/categories/payment/: Payment Infrastructure — 6 tools
- https://agent.c8.fit/categories/mcp/: MCP & Tooling — 9 tools
- https://agent.c8.fit/categories/memory/: Memory & State — 7 tools
- https://agent.c8.fit/compare/langgraph-vs-crewai/: LangGraph vs CrewAI comparison

## Benchmarks
- https://agent.c8.fit/benchmarks/: what each agent benchmark measures, its 2026 frontier, human baseline and credibility caveats
- https://agent.c8.fit/benchmarks/myth-langgraph-crewai-96: a widely circulated framework statistic that has no traceable source

## Dynamic pages (updated as events happen)
- https://agent.c8.fit/news.md: agent-related launches, funding, security incidents and framework changes, each with a date anchor, a source link and a verification status
- https://agent.c8.fit/news/: the same digest as an HTML page with filters

## Notes
- The Chinese version of every page is served under the /zh/ prefix.
- Every tool page also has a Markdown rendition at /tools/{slug}.md (e.g. /tools/langgraph.md).
- Star counts are point-in-time snapshots and are labelled with their as-of date.
- Some entries carry an explicit caveat where a widely-repeated claim could not be verified.