act101

act101 is a revolutionary dev tool that is taking the AI coding world by storm. It is the first and only tool that gives AI agents the ability to actually refactor and port code, not just chat about it. Think of it as a superpowered, language-aware operating system for your codebase that your AI agent can directly control. Built as a native Rust binary, act101 exposes a massive suite of 183 AST-aware refactor operations, 30 codebase analyzers, and 8 porting operations to popular AI coding agents like Claude Code, Cursor, Codex, and OpenCode via the Model Context Protocol (MCP). This means your AI agent can now do real, structural code work: extract a function from Python, rename a symbol across an entire Rust workspace, or port a C library to Rust with a structured state machine. With support for 163 programming languages, it is the most comprehensive code manipulation engine ever built. The best part? It is free for personal use, runs entirely on your machine with zero telemetry, and requires no indexing or caching, ensuring your code stays private and results are never stale. The developer community is buzzing about how this tool eliminates the "whole-file rewrite" problem, saving an average of 85% on tokens compared to file-based operations. act101 is not just a tool; it is the missing link that finally makes AI agents capable of serious, production-grade software engineering.

183 AST-Aware Refactor Operations

This is the core magic. act101 provides your AI agent with 183 typed, AST-aware operations that go far beyond simple find-and-replace. From common refactors like extract-function, rename, move-symbol, and inline, to advanced ones like convert-to-dataclass, extract-trait, and organize-imports, each operation understands the syntax tree of your code. This means the agent can make precise, cross-file changes that preserve comments, formatting, and logical correctness. Every operation is automatically checkpointed, allowing for instant undo if anything goes wrong, a feature that is a lifesaver for complex refactoring sessions.

8 Structured Porting Operations for Cross-Language Migration

Porting code between languages has historically been a nightmare, but act101 turns it into a manageable, state-machine-driven process. The tool introduces a four-step contract: port_contract to anchor the source-to-target migration, port_inventory to enumerate every symbol that must move, port_order to resolve dependency ordering, and a port_manifest state machine (init, add, update, remove, note) to track progress. This structured approach works between any two of the 163 supported grammars, making tasks like porting C to Rust or COBOL to Java a guided, agent-executed workflow rather than a chaotic rewrite.

30 Codebase Analyzers for Deep Structural Insight

Before your agent even touches a line of code, act101 can give it a complete structural map of your repository. The 30 built-in analyzers measure everything from cohesion and coupling to cycles, chokepoints, hotspots, dead code, layers, seams, and migration readiness. This allows the AI to understand the architecture, identify problem areas, and make informed decisions. It is like giving your agent x-ray vision into your codebase, enabling it to perform tasks like architecture audits and boundary analysis with unprecedented accuracy.

Native MCP Server with Zero Overhead

act101 is built as a single, native Rust binary that runs a built-in MCP server. This design choice has massive implications: there is no plugin runtime, no package graph, and no supply-chain attack surface. It requires no indexing or caching, meaning there is no cold start and results are never stale. It works instantly with Claude Code, Cursor, Windsurf, and any other MCP-aware client. The tool runs entirely on your machine with no telemetry, ensuring your proprietary code never leaves your local environment.

Agent-Executed Code Refactoring

Imagine asking your AI agent to "extract this logic into a function" or "rename this class across the entire project" and having it actually do it, correctly, with cross-file consistency. With act101, this is a reality. The agent can call operations like extract-function, move-symbol, and convert-to-dataclass on a Python codebase, or extract-trait on a Rust project, all while maintaining automatic checkpoints for instant undo. This eliminates the common frustration of AI agents doing whole-file rewrites that break formatting and lose comments.

Automated Language Migration

Porting a legacy C library to Rust, or a Ruby on Rails app to Elixir, is a monumental engineering task. act101 breaks this down into a structured, agent-executed workflow. The AI uses the port_contract to define the migration scope, port_inventory to list all symbols, and the port_manifest state machine to track each step. The agent can then systematically translate functions, data structures, and dependencies, with the tool ensuring that the migration is complete and consistent. This can cut migration timelines from months to days.

Deep Codebase Analysis and Health Checks

Before starting a major refactor or migration, you need to understand the current state of your code. With act101, you can ask your AI agent to perform a comprehensive analysis. The agent can run analyzers for coupling, cycles, dead code, hotspots, and migration readiness in seconds. This is invaluable for architecture audits, code reviews, and change impact analysis. The agent can then produce a structured report, identifying the most critical areas to address first, saving teams weeks of manual investigation.

Pre-Built Agent Skills for Common Workflows

act101 comes with 10 pre-built agent skills that compose lower-level operations into powerful, ready-to-use workflows. Skills like architecture-audit, code-review, refactoring, migration-assessment, and boundary-analysis can be invoked with a simple command like /skill-name in Claude Code or Cursor. This makes it incredibly easy for developers to leverage the full power of act101 without needing to learn every single operation. It is like having a team of expert AI engineers specialized in different tasks.

How does act101 ensure code safety during refactoring?

act101 employs automatic checkpointing on every single operation. Before any change is made, a checkpoint is created. If the result is not what you or your agent expected, you can instantly undo the operation, restoring the code to its previous state. This safety net is crucial for complex, multi-file refactors and is a major reason why developers trust act101 for production code.

Does act101 require my code to be uploaded to a cloud server?

No. act101 is a native binary that runs entirely on your local machine. Your code never leaves your computer. There is no telemetry, and the only time it contacts our servers is for license verification. This design ensures that your proprietary and sensitive code remains completely private and secure.

What languages does act101 support?

act101 supports an incredible 163 programming languages, all packed into a single Rust binary. This includes all major languages like Python, Rust, TypeScript, Go, Java, C, C++, Ruby, Elixir, COBOL, and many more. The tool uses tree-sitter grammars for parsing, ensuring broad and accurate syntax tree coverage for both refactoring and porting operations.

How does act101 integrate with my existing AI coding tools?

act101 is built around the Model Context Protocol (MCP), making it a first-class citizen for all MCP-aware clients. It works out of the box with Claude Code, Cursor, Codex, OpenCode, and Windsurf. You simply install the binary and configure your AI tool to connect to the built-in MCP server. Once connected, your AI agent gains access to all 183 refactor operations, 30 analyzers, and 8 porting operations.

act101 is free for personal use. This includes access to all 163 grammars, 183 AST refactor operations, 30 codebase analyzers, 8 porting operations, and 10 pre-built agent skills. For commercial and team usage, please see the pricing page on the website for available plans and tiers.