Claktor
Claktor is a refactoring CLI for coding agents. It finds repeated logic—even when names and syntax differ—and measures how much smaller a behavior-preserving refactor makes production code, so teams and agents have less code to read, test, secure, and maintain.
At a Glance
- Software development teams using coding agents
- Individual developers and coding-agent users
- Open-source repository maintainers
- Teams running agent workflows in CI
- +1 more
AI Tools by Claktor
(1)Claktor
Refactoring CLI for Coding Agents
Discussions
No discussions yet
Be the first to start a discussion about Claktor
Latest News
Mateusz Bednarski announced the public launch of Claktor, a CLI for coding agents that finds repeated logic and measures refactoring reductions.
Show HN: Claktor – a refactoring CLI for coding agents.
Why do coding agents repeat logic that already exists? Claktor explains the context-window and duplication problem and its fuzzy AST approach.
How do you use Claktor? Guidance for on-demand, sub-agent, and scheduled-CI workflows.
Products & Services
A command-line tool that coding agents run to map a repository, identify repeated logic using parser/fuzzy-AST analysis, rank opportunities by estimated savings, and measure normalized-token reductions after a refactor. It supports TypeScript, JavaScript, Python, and Rust and runs on Linux and macOS on x86_64 and arm64.
Market Position
Claktor positions itself as an agent-oriented refactoring CLI rather than a conventional style checker: its differentiators are fuzzy structural matching for non-identical repeated logic, ranking by estimated savings, and a normalized-token metric intended to approximate the future reading and maintenance cost for both agents and engineers.
Leadership
Founders
Mateusz Bednarski
Claktor co-founder and builder. He studied Computer Science at King's College London (2019–2022), worked on crypto infrastructure and DeFi arbitrage at Eisler Capital, built trading systems and cross-chain/DeFi projects, and built and ran a CoW Protocol solver at Copium Capital that routed $15B in user orders. He was also a co-founder of QVote and tech lead at Ignite DAO.
Executive Team
Mateusz Bednarski
Co-Founder
Computer Science graduate of King's College London; previously worked in crypto infrastructure and DeFi arbitrage at Eisler Capital, built trading systems and a CoW Protocol solver, and co-founded QVote.
Founding Story
Bednarski says he started Claktor after becoming frustrated with the code smells and duplicated logic that AI coding agents produced while writing Rust. The initial idea evolved from an opinionated code-quality detector into a tool that uses custom parsing/fuzzy AST matching to rank objectively useful deduplication opportunities by normalized-token savings.
Business Model
Revenue Model
The current CLI binary is distributed through an install script and GitHub releases under the PolyForm Shield License. Claktor's agent documentation says a paid version with more features and refactoring workflows is in development; public pricing is not specified.
Target Markets
- Software development teams using coding agents
- Individual developers and coding-agent users
- Open-source repository maintainers
- Teams running agent workflows in CI
- Projects written in TypeScript, JavaScript, Python, or Rust
- Reducing duplicated logic in repositories maintained or extended by AI coding agents
- Running periodic refactoring on demand, in a sub-agent, or in scheduled CI
- Creating focused pull requests that consolidate repeated implementations while preserving behavior
- Refactoring TypeScript, JavaScript, Python, and Rust production code
- steipete/Markdansi — public case study in which three streaming flush helpers were consolidated into one implementation, removing 50 production lines and about 180 normalized tokens