LuciferMornens

LuciferMornens / kontext-engine

Public

Stop feeding your AI agent the entire codebase. Give it ctx instead. FIrst context engine for AI agents that actually works.

53
0
100% credibility
Found Feb 17, 2026 at 27 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
TypeScript
AI Summary

A local tool that scans code projects to enable fast, intelligent searches for relevant code snippets, perfect for feeding context to AI coding assistants.

How It Works

1
📖 Discover ctx

You hear about ctx, a friendly tool that helps AI assistants deeply understand your entire code project without reading everything.

2
🛠️ Get ctx set up

With one easy step, you add ctx to your computer so it's ready to use anywhere.

3
🏗️ Teach ctx your code

You go to your project folder and tell ctx to learn your code files, which takes just seconds.

4
🔍 Ask and discover

You type a question like 'how does login work?' and instantly see the exact code spots that answer it, ranked perfectly.

5
👀 Keep it fresh

You turn on auto-watch so ctx updates itself whenever you change your code.

🎉 AI gets super smart

Now your AI helper knows your project inside out, giving precise help every time you code.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 27 to 53 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is kontext-engine?

Kontext-engine is a TypeScript CLI tool that indexes your codebase into a local SQLite database, letting AI agents query precise file ranges and snippets instead of dumping the whole repo. Run `ctx init` to index (5s for 1K files), then `ctx query "auth middleware"` for fused results from semantic vectors, full-text, AST symbols, paths, and deps—all in agent-ready JSON. Stop feeding AI the entire codebase; kontext engineering is the new prompt engineering.

Why is it gaining traction?

It fuses five search strategies with Reciprocal Rank Fusion for millisecond results that beat grep or naive RAG, with smart re-ranking that deprioritizes imports/tests and boosts paths/exports. 100% local embeddings via ONNX, watch mode for live updates, and `ctx ask` for LLM-steered natural queries make it dead simple—no servers, no plugins. Developers love the bash integration for tools like Cursor or Claude.

Who should use this?

AI agent builders scripting shell calls for targeted context. Backend devs tracing deps in monorepos. Anyone tired of context window blowups in IDE agents or custom RAG pipelines.

Verdict

Try it for agent workflows—solid docs, CLI, and tests make the 43 stars and 1.0% credibility score forgivable for early-stage TypeScript. Maturity lags (some commands WIP), but incremental indexing and local-first design deliver immediate wins over bloated alternatives.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.