The index of AI agent tooling,
curated by people who use it.
Every coding agent, framework, and eval tool, verified against its docs and repo, with novelty-scored knowledge extracted from curated sources. No vendor self-ranking. No stale listicles.
50 tools tracked · 35 open source · 11 curated insights · refreshed weekly
Built to be the source you can trust
Practitioner-verified
Each listing carries a verified date and a named curator. Specs are checked against the tool's docs and repo, not scraped from a marketing page.
Novelty-scored knowledge
Insights from curated videos and articles are scored by how much new signal they add. We surface what is genuinely different, not what every review already says.
Open data in git
The dataset is plain YAML and JSON in the repo. Every change is a commit, so you can see and trust exactly what changed, and when.
Weekly freshness
A scheduled action refreshes GitHub stars, recent activity, and the knowledge graph every week, so listings never quietly go stale.
Three maps of the agent stack
Coding agents
AI tools that write, edit, and navigate code, from IDE assistants to autonomous terminal agents.
10 toolsAgent frameworks
Libraries and SDKs for building, orchestrating, and running AI agents in your own applications.
19 toolsEval and observability
Tracing, evaluation, and monitoring platforms so you can ship agents you actually trust.
Verified picks to start with
Aider
AI pair programming in your terminal, git-native.
Claude Code
Anthropic's agentic CLI coding assistant.
Cursor
The AI-first code editor built on VS Code.
Langfuse
The most-used open-source LLM observability tool.
OpenCode
A model-agnostic, local-first coding agent TUI.
A knowledge layer, not just a list
We ingest transcripts from the best practitioner videos, extract the claims that matter, and score each one for novelty against the rest. The result is a living graph that connects tools to the concepts and tradeoffs that define them.
“Persistent memory is the biggest multiplier: a directory that logs corrections, read at session start, means every fix compounds.”
novelty 0.92 · sourced from practitioner notes