Navegación basada en árboles, refactorización semántica, análisis de código base y herramientas de portabilidad de lenguajes para agentes de codificación que admiten 163 gramáticas.
act101 is a developer tool that fundamentally changes how AI coding agents interact with source code. Unlike conventional tools that treat code as plain text, act101 gives agents language-aware capabilities through a native Rust binary that understands 163 programming grammars. It operates as a Model Context Protocol (MCP) server, exposing 183 AST refactor operations, 30 codebase analyzers, 15 query operations, and 8 porting operations — all running locally with no telemetry, no indexing, and no cold starts. Your code never leaves your machine.
The agent can call extract-function, rename, move-symbol, inline, convert-to-dataclass, extract-trait, add-type-hints, generate-init, organize-imports, and 174 more — all across the entire codebase with cross-file consistency. Every operation creates an automatic checkpoint, and the agent can undo instantly if the result looks wrong.
Porting follows a four-step state machine: port_contract anchors the source-to-target migration, port_inventory enumerates every symbol that must move, port_order resolves dependency ordering, and port_manifest tracks progress through init/add/update/remove/note states. This works between any two of the 163 supported grammars.
The analyzers measure cohesion, coupling, cycles, chokepoints, hotspots, dead code, layers, seams, clusters, fan balance, migration readiness, and type completeness. The pre-built skills — architecture-audit, code-review, refactoring, code-navigation, code-generation, migration-assessment, boundary-analysis, change-impact, health-check, and architectural-refactoring — compose these operations into common engineering workflows.
act is the first dev tool that lets an AI agent actually do language-aware code work — not just chat about it.
This is the critical distinction. Before act, AI agents could only perform whole-file rewrites that lost comments, broke formatting, and offered no undo. act gives agents typed, AST-aware operations that preserve code structure, maintain cross-file consistency, and provide instant rollback. The single native Rust binary has no plugin runtime, no package graph, and no supply-chain attack surface — it parses on call and returns results that are never stale.
You're building with AI coding agents and need them to perform real refactoring — not just generate text. You're planning a language migration and want a structured, agent-driven approach that preserves code semantics. You're analyzing large codebases for coupling, dead code, or architectural boundaries and want results without indexing or caching delays. act is free for personal use and works with any MCP-aware client.
Other tools you might consider
Loading comments…
