Agency is an open-source platform developed at the UC Berkeley Sky Computing Lab for running, debugging, profiling and controlling AI agents. As agents tackle longer tasks, invoke tools and work in parallel, model traces only tell part of the story, so Agency focuses on questions like what an agent is waiting for and where time is being spent.
Be the first to rate Agency. Your feedback helps other users decide.
Systems profiling shows where time goes, beyond what a model trace records.
The platform runs, debugs, profiles and controls agents in one place.
It is built for agents that work in parallel across longer tasks.
Similar tools in the same category — compare side-by-side
Renor is memory for coding agents: it keeps what you tell Claude Code, Codex or Cursor in a folder o...
catnap is a Claude Code mod that lets Claude compact its own conversation and carry on by itself. Cl...
Pando deploys and shares apps. It turns the apps your team builds with AI into tools everyone can us...
Guides and comparisons about code tools
Ten launches aimed at the same bottleneck: agents re-read their own history, and every re-read is billed. Compression, code mode, prices and measurement.
Eight new tools for what sits under an agent: deterministic memory, an agentic database, distributed inference, hardware checks and spend guards.
Nine new tools that put something deterministic between an agent and what it touches: gateways, policy firewalls, AST breakers, SQL guards and usage ceilings.
Nine new tools for evidence rather than dashboards: signed records, decision logs, MCP analytics, protocol tests, sandboxes and long-running agent supervision.
Nine new tools aimed at the same wall from different angles: indexed wikis, stale memory, knowledge graphs, reversible compression and multi-agent rooms.
