Skip to content
View tensorninja's full-sized avatar

Block or report tensorninja

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
tensorninja/README.md

👋 Hi I'm Thorsten

I design data platforms and the AI systems that run on them. I live in Aachen, Germany.

thorstenborn.com · caudra.ai · LinkedIn

Caudra

Caudra is the complete coding agent for your terminal. I created it and maintain it, and I use it for my own work every day.

Caudra in a terminal. You type guidance into the chat of a running subagent while it patches src/pagination.test.ts, and the guidance waits in the task queue.

I kept adding plugins to get the features I wanted. Often I could not build the plugin I needed, because the agent had no hook in the place it needed one. So I built those features into the agent itself, where they can work together.

Caudra is written in Rust and runs on macOS, Linux, and Windows. It is an independent fork of Maki by Tony Solomonik.

curl -fsSL https://caudra.ai/install.sh | sh

ninfer-4090

As a hobby I maintain ninfer-4090. It is my port of Don-Chad's NInfer-3090, a fork of Neroued's NInfer engine, to the RTX 4090. It runs Qwen3.8-27B on one 24 GB card, and Caudra uses it as a local model and decision engine.

Pinned Loading

  1. caudra/caudra caudra/caudra Public

    Caudra is growing from a terminal coding agent into a complete AI workspace: a cockpit to run agents, persistent agents with memory, and the TUI agent you can use today. Optional remote machines an…

    Rust 11 1

  2. ninfer-4090 ninfer-4090 Public

    Forked from sergiuszm/ninfer-4090

    From-scratch C++/CUDA engine for Qwen3.8-27B on one RTX 4090: 3,600 tok/s prefill, 149 tok/s MTP decode, full 262K context. Works with any coding agent over OpenAI and Anthropic APIs, best with Cau…

    C++ 16 6