v0.53 · May 2026

A DeepSeek-native
coding agent,
for your terminal.

Reasonix talks straight to api.deepseek.com. The loop is append-only, engineered around DeepSeek's byte-stable prefix cache — long sessions hold 90%+ cache hit and input-token cost collapses to ~1/5. Terminal-first, leave it running.

Get started → View the GitHub repo
94% Cache Hit
2.5× Cost Down
MIT License
11k+ GitHub Stars
Overview

What makes Reasonix different from a normal coding agent

01

A real feedback loop, not just automation

The agent proposes a change, runs the experiment, reads the result, and keeps only the ideas that improve the metric.

02

One editable file keeps the system legible

Most research decisions stay inside train.py, which makes each iteration easy to inspect and compare.

03

A fixed budget keeps comparisons honest

Every run gets the same five-minute wall-clock window, so changes compete under the same constraint.

04

Engineered around prefix-cache stability

Cache stability isn't a feature you turn on; it's an invariant the loop is designed around. Every layer is tuned to the byte-stable prefix-cache mechanic.

§01 · Install

Two steps. No global install.

Node ≥ 22 on macOS / Linux / Windows (PowerShell · Git Bash · Windows Terminal). The first launch walks you through pasting a DeepSeek API key.

01

Get a DeepSeek API key

Head to the DeepSeek platform and create a key. Pay-as-you-go; cached tokens bill at 1/5 of the regular rate.

$0.07 /Mtok in · $0.014 /Mtok cached
02

Enter your project

$ cd /path/to/my-project

Tools sandboxed to launch directory.

03

Launch the TUI

$ npx reasonix code

First-launch wizard wires up the key. Or run it without installing:

$ cd my-project && npx reasonix code
§02 · Three Pillars

Why DeepSeek-native

Reasonix only targets DeepSeek because the loop's invariants are designed against DeepSeek's cache mechanics. Same model, same API — what changes is the engineering stance of the loop.

Cache-First Loop

A byte-stable loop. DeepSeek's prefix cache fingerprints prompts from byte 0. The Reasonix loop is append-only — no reordering, no marker-based compaction — so the cached prefix survives every tool call.

94% cache hit · long sessions

R1 Thought Harvest

Reasoning-chain harvest. When DeepSeek's R1 reasoning block contains escaped tool calls, Reasonix scavenges and re-routes them — no turn wasted.

Tool-Call Repair

Schema-aware self-heal. When tool calls drift from the expected schema, Reasonix repairs them inline instead of discarding the turn — keeps the prefix intact.

§03 · Features

Engineering stance around the DeepSeek API.

A dozen-odd tools combine into what looks like a simple CLI — but every layer beneath is working to keep cache hit, cost, and stability where they need to be.

F-01

Terminal-native TUI

Not another IDE plugin. git diff handles diffs, ls handles file trees — your terminal is the workspace. Built with TypeScript + Ink.

F-02

V4 two-tier

V4-Flash by default for cheap iteration; /pro lifts a single turn to V4-Pro; /preset max makes the whole session run on Pro.

F-03

MCP first-class

One line — --mcp "name=cmd args" — and an external server is wired in; its tools merge into the same registry under a prefix. Supports stdio, SSE, Streamable HTTP.

F-04

Sandbox + plan gate

Every built-in tool is sandboxed to the launch dir; /plan puts the session behind a read-only audit gate — no writes until the plan is approved.

F-05

Composable skills

Drop a Markdown file in .reasonix/skills/*.md; frontmatter supports runAs: subagent and allowed-tools for isolated execution.

F-06

Replay & events

Every event hits disk — replay any past session, run stats on token/cache/cost, audit your loop's behaviour. reasonix replay / events / stats.

How it works

The loop stays small so each improvement is easy to judge

One editable training file, one fixed wall-clock budget, and one metric keep the system narrow enough to inspect and strong enough to iterate.

01

Inspect the setup

The agent reads the instructions, understands the experiment target, and decides what to change.

02

Edit one file

Architecture, optimizer, hyperparameters, and training logic all live in the same training file.

03

Train for five minutes

Every experiment gets the same wall-clock budget, which makes short-run results more directly comparable.

04

Keep or discard

The run is scored on validation bits per byte. Better changes stay. Worse changes get thrown away.

How it compares

Reasonix vs. other coding agents

Reasonix Claude Code Cursor Aider
Backend DeepSeek Anthropic OpenAI / Anthropic any (OpenRouter)
License MIT closed closed Apache 2
Cost profile low per task premium subscription + use varies
DeepSeek prefix-cache engineered not applicable not applicable incidental
Embedded dashboard yes n/a (IDE)
Plan mode · MCP · hooks · skills yes yes yes partial
Open community yes yes
FAQ

The fastest answers to the questions people ask first

Why DeepSeek only? Can I swap to Claude / GPT?

It's a design choice, not a limitation. DeepSeek's prefix cache fingerprints prompts from byte 0; the Reasonix loop is engineered around that invariant — long sessions hold ~94% cache hit. Coupling to one backend is the feature.

Is it free?

Reasonix itself is MIT-licensed and free. The DeepSeek API key is paid: V4-Flash is $0.07/Mtok uncached and $0.014/Mtok cached. In long sessions the bill typically lands at ~1/3 of comparable generic tooling.

Will there be an IDE plugin?

No. Reasonix is terminal-first. git diff does diffs; ls does file trees. The desktop is a visual companion to the CLI, not a Cursor replacement.

Can I point it at a self-hosted DeepSeek endpoint?

Yes. Since 0.30 we accept non-standard key prefixes for self-hosted DeepSeek endpoints. Just point baseUrl at your internal address — the loop, cache strategy, and tool protocol are unchanged.

How do I write my own skill?

No remote registry — just write a file. Inside the TUI run /skill new my-skill to scaffold a project-local skill; add --global to put it under ~/.reasonix/skills for reuse across projects.

What do you need to run it?

Node ≥ 22 on macOS, Linux, or Windows. A DeepSeek API key. That's it. The desktop client bundles its own Node runtime.

Primary sources

Every claim on this page is grounded in the repo README or linked public discussion

GitHub Repo

The source for the README, project structure, and quick-start commands.

github.com/esengine/DeepSeek-Reasonix →

Official Website

Getting started, dashboard mockup, TUI mockup, and full documentation.

esengine.github.io/DeepSeek-Reasonix →

Architecture Docs

Three pillars: cache-first loop, tool-call repair, cost control.

ARCHITECTURE.md →