act101

act101 is a developer tool that gives AI coding agents structural, AST-aware operations for semantic refactoring and language migration. Rather than rewriting files as plain text, it exposes precise edits so formatting, comments, imports and project layout stay intact, and every change can be checkpointed and undone.

The platform supports 163 programming grammars and offers more than 180 refactoring operations, from extracting functions and renaming symbols to moving modules and reorganising imports. It runs as an MCP-native server so agents like Claude Code, Cursor, Codex and OpenCode can call into it and perform real code transformations instead of approximate text edits.

act101 also targets large-scale migrations and analysis. It provides structured migration workflows—using contracts, inventories, dependency ordering, manifests and migration state tracking—to reduce behavioural drift and dependency problems during conversions such as C to Rust or Python 2 to Python 3. There are over 30 repository analysis tools for coupling, dead code, hotspots, dependency cycles, type completeness and migration readiness, which AI agents can use before making changes.

Security and local control are emphasised: act101 is a single native Rust binary with no plugin runtime and no cloud uploads, parsing projects on demand without indexing or caching. It’s optimised for token efficiency too, returning AST-derived slices of code so agents don’t need to load whole files into their context windows—benchmarks claim some operations cut token usage by over 95%.

• Official site: https://act101.ai/
• 163 grammars and 180+ refactoring operations
• Works with Claude Code, Cursor, Codex, OpenCode (MCP-native)
• Structured migration workflows and 30+ analysis tools
• Single Rust binary, local execution, checkpointing and undo