Experience

rustkit-ai

Open Source
AI
Rust

Open source organization for Rust developer tools used in AI-assisted work. Main project: semtree, semantic code search that runs locally by default, published on crates.io under MIT.

Language
Rust
AI/ML
RAGEmbeddingsTree-sitter
Protocol
MCP

rustkit-ai: Rust tools for AI-assisted development

rustkit-ai is an open source organization of Rust tools for AI-assisted development, which I co-founded with Gokhan Kabar. Everything is MIT licensed and published on crates.io.

The tooling was born out of practical needs met during internal R&D work at Ippon Technologies, where the same problems (context loss, token costs, understanding a large codebase) kept coming back on real production code.

semtree

semtree is my main project, which I built and maintain: semantic code search that runs locally by default.

  • Parses a codebase with tree-sitter (20 languages) and chunks it along the syntax tree, so a chunk is a function or a class rather than an arbitrary window of lines.
  • Embeds the chunks on your machine with fastembed and stores them in an HNSW index next to the repository.
  • Ranks results by fusing vector similarity with BM25 (hybrid by default), with semantic-only and lexical-only modes available.
  • Keeps the embedder and the vector store behind traits, so a local Ollama model or a remote API can replace the defaults.
  • Ships as a library, a CLI and an MCP server for coding agents, split into 9 crates on crates.io.
cargo install semtree-cli

My other tools

trimcp MCP proxy that compresses and caches MCP tool output before it reaches the model. Auto-detects Claude Code and Cursor MCP configurations.

cargo install trimcp
trimcp setup

semstore, co-written with Gokhan Local semantic search for Rust applications: store text, search by meaning, no cloud API required.

Also in the organization

Gokhan Kabar builds aimemo (persistent memory for AI coding agents), tersify (strips noise from code and text before it goes into an LLM context) and mcpkill (a semantic cache proxy for MCP servers).

References

Related experience

Ippon Technologies

Designed and built an internal AI platform for legacy modernization at Ippon Technologies: a Rust CLI, a web app and an MCP server on one core, rolled out to 120+ consultants.

RustTypeScriptAIMCP

Related articles

MCP architecture: Rust CLI connected to an IDE via an MCP server

Model Context Protocol (MCP): Exposing a Rust CLI to Your IDE

A practical deep-dive on using MCP to expose a Rust CLI to an AI agent in the IDE: architecture, best practices, and TypeScript implementation.

AISoftware Engineering
A terminal session running an AI coding agent alongside git, shell, and custom CLI tools

Why I Left the AI IDE for the Terminal, and Built Tooling Around It

I traded my AI-enhanced IDE for a terminal agent. It stuck, not because the terminal is cooler, but because it's composable. Here's what changed, the honest trade-offs, and the tooling I had to build.

AISoftware Engineering
AWS EC2 infrastructure for a self-hosted Rust RAG app with Terraform

Deploying a Self-Hosted RAG App on AWS with Terraform: Why I Chose EC2 Over Bedrock

A practical breakdown of deploying a Rust RAG application on AWS EC2 using Terraform: stateful Qdrant data on EBS, GitLab OIDC for credential-free CI/CD, and the real trade-offs between self-hosting and AWS Bedrock.

DevOpsAWSRust