DeepSeek-Reasonix
SHA-256A DeepSeek-native AI coding agent for your terminal. Config- and plugin-driven, tuned for prefix-cache stability to keep token costs low across long sessions.
Smart Download
Download Download Version
v1.11.0 · 7.8 MB
A DeepSeek-native terminal AI coding agent with prefix-cache optimization, config-driven and plugin-extensible.
Core Features
- Config-driven: All providers, models, tools, and plugins declared in a single TOML file; no hardcoded models
- Multi-model composable: Run executor + planner models in separate cache-stable sessions
- Plugin-driven: Built-in tools self-register at compile time; external tools integrate via MCP-compatible stdio JSON-RPC
- Zero-friction distribution: Single static Go binary, cross-compiled for 6 targets, no runtime dependencies
What It Can't Do
- •Only supports DeepSeek or OpenAI-compatible endpoints; no local model support (e.g., Ollama). API keys are referenced via environment variables, never written in plain text to config files. Windows builds are code-signed; macOS may prompt to allow unsigned binary.
Use Cases
- Long coding sessions: Leverages prefix cache to reduce token costs during extensive refactoring or large project maintenance
- Team collaboration: Unify toolchain via config files, support multi-model workflow for improved efficiency
Detailed Introduction
Reasonix is a Go-based, single-binary AI coding agent designed specifically for DeepSeek models. It leverages DeepSeek's prefix cache to dramatically reduce token costs during long-running coding sessions. Unlike most AI coding assistants that are cloud-dependent or require heavy runtimes, Reasonix runs entirely in your terminal with zero external dependencies (just a TOML parser). It features a config-driven architecture where providers, agents, tools, and plugins are all declared in a single TOML file. Built-in tools self-register at compile time, and external tools integrate via MCP-compatible subprocess JSON-RPC. Reasonix supports multi-model setups (executor + planner) and can be installed via npm, Homebrew, or built from source. It is fully open source under MIT license.
Tags
Getting Started
Download installer
Click the button above to download the installer for your system
Install the software
Double-click the downloaded installer and follow the prompts
Install: npm i -g reasonix (recommended) or brew install esengine/reasonix/reasonix (macOS)
Initialize: Run reasonix setup to generate reasonix.toml via interactive wizard
Set API key: export DEEPSEEK_API_KEY=sk-... or let the wizard save it to the OS credential store
- Install: npm i -g reasonix (recommended) or brew install esengine/reasonix/reasonix (macOS)
- Initialize: Run reasonix setup to generate reasonix.toml via interactive wizard
- Set API key: export DEEPSEEK_API_KEY=sk-... or let the wizard save it to the OS credential store
SHA-256 checksum verified
Checksum extracted from GitHub official Release page
SHA256 Checksum
f821cee0579979224f8ccf553a61a429f81929aeb4838cf9d57b37f880817403This checksum is extracted from the GitHub Release page. Verify file integrity after download.
All SHA-256 checksums on this platform are extracted from the project's official GitHub Release page, without any modification. You can independently verify them on the GitHub Releases page.
Open Source Transparency
View GitHub SourceUninstall Info
npm uninstall -g reasonix (npm); brew uninstall reasonix (Homebrew); or simply delete the binary and ~/.reasonix/ config folder
No Extra Dependencies
Ready to use after download. No additional runtime required.
Similar Projects
Visual Studio Code (Code - OSS)
A free, open-source code editor by Microsoft with rich extensions, debugging, and Git integration. The most popular developer tool worldwide.
scrcpy
Display and control your Android device from your computer via USB or TCP/IP. No root required, high performance, low latency, supports audio and video mirroring.
godot
Godot Engine is a free, open-source, cross-platform 2D and 3D game engine with a unified editor. Unlike Unity or Unreal, it's entirely MIT-licensed and community-driven, offering a lightweight yet powerful node-based workflow that runs on a single executable.