Thirty-three more agents, seven kinds, and a duel you can actually finish
The board grows to 74 agents including frameworks and cloud agents, every critic now argues from different facts, duels play through eleven matchups, and the site gained a light theme and category pages.
El Amigo writes the lead. Day two, and the board is nearly twice the size: 74 agents, sorted into seven kinds so you can start from the one you would actually install. New today: Google's Antigravity, OpenAI's Codex cloud agent, the mini-swe-agent successor to SWE-agent, Mistral Vibe, Qwen Code, and JetBrains AI Assistant.
Frameworks are on the board now. If you are building the agent rather than buying one, LangGraph, CrewAI, the OpenAI Agents SDK and the Claude Agent SDK have rows and reviews. La Jefa treats them as platform dependencies, which is the right lens.
The panel argues from different facts now. Yesterday the seven critics too often quoted the same price or the same sandbox line. Each critic now owns one question and cites facts no other critic uses, and a check enforces it across every pair of reviews on every agent.
Four sunsets in one day. GitHub Spark is deprecated, Firebase Studio has a shutdown date, Supermaven is sunset after its acquisition, and Codegen folded into ClickUp. They stay listed and labeled, because people still search for them.
Duels play through. Vote, see the community split and which critics agree with you, move to the next. Eleven matchups so far, from junior's first week to sixty-seat procurement. Your picks are saved to your account, and the page tells you which critic you think like.
Tomorrow: the pipeline begins pulling releases and pricing changes for all 74 rows.