> [!IMPORTANT]
> **Reasonix 1.0 is a ground-up rewrite in Go** — this branch (`main-v2`) is the new default and where development happens now.
> The earlier `0.x` TypeScript releases are **legacy**, living on the [`v1`](https://github.com/esengine/DeepSeek-Reasonix/tree/v1) branch (maintenance only).
> See the **[migration guide](./docs/MIGRATING.md)**. `npm i -g reasonix` stays the install command — `1.0.0`+ delivers the Go binary, `0.x` is the legacy TS build.
A DeepSeek-native AI coding agent for your terminal.
A config- and plugin-driven harness — a single static Go binary, tuned around DeepSeek's prefix cache so token costs stay low across long sessions.
> [!IMPORTANT]
> **Community · 加入社区** — bilingual Discord for setup help (`#help` / `#求助`), workflow showcases, and feature ideas. → ****
## Features
- **Config-driven.** Providers, the agent, enabled tools, and plugins are all
declared in `reasonix.toml`. No hardcoded models.
- **Multi-model & composable.** DeepSeek (flash/pro) and MiMo ship as presets;
any OpenAI-compatible endpoint is a config entry, not new code. Optionally run
two models together (executor + planner) in separate, cache-stable sessions.
- **Plugin-driven.** External tools run as subprocesses over stdio JSON-RPC
(MCP-compatible). Built-in tools self-register at compile time.
- **Zero-friction distribution.** `CGO_ENABLED=0` single binary; cross-compile
to six targets with one command. The only dependency is a TOML parser.
- **Math research tools.** Built-in **[Iteris](#iteris--agentic-research-pipeline)** agentic
research pipeline (4-phase: Foundation → Experiment → Proof → Review) with
desktop visualization, plus **[Matlas](#matlas--semantic-math-search)** semantic
search over 8M+ peer-reviewed mathematical statements.
## Install
```sh
npm i -g reasonix # any OS; pulls the prebuilt native binary
brew install esengine/reasonix/reasonix # macOS
```
Prebuilt archives (`darwin|linux|windows × amd64|arm64`) and `SHA256SUMS` are on
every [GitHub release](https://github.com/esengine/DeepSeek-Reasonix/releases).
### Code signing
Windows builds are code-signed with a free certificate provided by the
[SignPath Foundation](https://signpath.org/), with signing through
[SignPath.io](https://signpath.io/).
### Build from source
```sh
make build # -> bin/reasonix(.exe)
make cross # -> dist/ (darwin|linux|windows × amd64|arm64)
```
## Quick start
```sh
reasonix setup # config wizard → ./reasonix.toml
export DEEPSEEK_API_KEY=sk-... # or put it in .env (see .env.example)
reasonix chat # then run /init to generate AGENTS.md (project memory)
reasonix run "implement the TODOs in main.go"
reasonix run --model mimo-pro "add unit tests for this function"
echo "explain this code" | reasonix run
```
## Configuration
A minimal `reasonix.toml` — one provider and a default model — is enough to start:
```toml
default_model = "deepseek-flash"
[[providers]]
name = "deepseek-flash"
kind = "openai"
base_url = "https://api.deepseek.com"
model = "deepseek-v4-flash"
api_key_env = "DEEPSEEK_API_KEY"
```
Resolution order is **flag > `./reasonix.toml` > the user config file >
built-in defaults**; the user file lives in your OS config dir — `~/.config/reasonix/`
on Linux, `~/Library/Application Support/reasonix/` on macOS, `%AppData%\reasonix\` on
Windows. Secrets come from the environment via `api_key_env` and are
never written to config files. Permissions, the sandbox, plugins (MCP), slash
commands, `@` references, and two-model setup are all in the
**[Guide](./docs/GUIDE.md)**.
## Documentation
- **[Guide](./docs/GUIDE.md)** — configuration, permissions & sandbox, plugins
(MCP), slash commands, `@` references, two-model collaboration.
- **[Bot guide](./docs/BOT_GUIDE.md)** — connect Feishu, Lark, and WeChat bots
from the desktop app, then use approvals, YOLO, and commands from IM.
- **[Spec](./docs/SPEC.md)** — engineering contract: architecture, registries,
data types, and roadmap.
- **[Migrating from 0.x](./docs/MIGRATING.md)** — moving from the legacy
TypeScript releases to the 1.0 Go rewrite.
- **[Checkpoints & rewind](./docs/CHECKPOINTS.md)** — the snapshot-based edit
safety net (Esc-Esc / `/rewind`).
- **[Iteris pipeline](#iteris--agentic-research-pipeline)** — automated 4-phase
research loop for computational mathematics, with desktop visualization.
- **[Matlas search](#matlas--semantic-math-search)** — semantic search across
8.07M peer-reviewed mathematical statements (180 journals, 1826–present).
## Computational Math Research
Reasonix ships with tools for computational mathematics research, powered by
the **[Iteris](https://arxiv.org/abs/2606.02484)** agentic research loop and
**[Matlas](https://arxiv.org/abs/2604.17484)** semantic math search.
### Iteris — Agentic Research Pipeline
A 4-phase research loop for computational mathematics, with dedicated desktop
visualization:
```
Foundation → Experiment → Proof → Review
```
**Quick start in the desktop app:**
```sh
# Phase 1: Audit sources and freeze definitions
/iteris-foundation Analyze the spectral radius bound for M⁻¹A
# Phase 2: Run numerical experiments (evidence is NEVER proof)
/iteris-experiment Test convergence for p=0.5, 1.0, 1.5
# Phase 3: Construct proof artifacts
/iteris-proof Prove the spectral radius bound ρ_n = O(n^{-1}) for p > 1
# Phase 4: Evaluate evidence quality
/iteris-review Review all artifacts under .reasonix/iteris/
```
**Pipeline visualization**: Open the desktop app, go to **Settings → Skills**,
click any `iteris-*` skill to open the **IterisPanel** — a 4-phase progress
bar with color-coded status, file preview tabs (PROBLEM.md, EXPERIMENT.md,
PROOF.md, REVIEW.md), and phase history timeline.
**CLI usage:**
```sh
reasonix run "/iteris-foundation freeze definitions for CG vs RCD comparison"
```
State persists under `.reasonix/iteris/` (PROBLEM.md, EXPERIMENT.md, PROOF.md,
REVIEW.md, TASK_POOL.json). Review phase enforces automated quality checks:
≥1 actionable finding required, max 2 restart cycles before forced fixable
verdict.
### Matlas — Semantic Math Search
Built-in `matlas_search` tool searches **8.07M peer-reviewed mathematical
statements** across 180 journals (1826–present) via the Matlas REST API.
**In the desktop composer** (any workspace):
```
/matlas spectral method convergence rate
```
The model calls `matlas_search` and returns ranked results with title, type
(paper/book), authors, journal, year, DOI, and the mathematical statement
text. Also triggerable with natural language:
> Find papers on Navier-Stokes regularity in Matlas
No API key or configuration required — the tool works out of the box as a
read-only built-in.
\n\n## Star History
## Support
If Reasonix has been useful and you'd like to say thanks, you can. It stays a coffee, not a contract — donations don't buy feature priority or change how issues get triaged.
- **International** — PayPal: [paypal.me/yuhuahui](https://paypal.me/yuhuahui)
- **国内** — 微信支付（扫码）
## Acknowledgments
A small list of folks whose work has shaped Reasonix the most — measured
by both commit count and code volume. **Listed alphabetically, no ordering
of importance.** The full contributor graph is on
[GitHub](https://github.com/esengine/DeepSeek-Reasonix/graphs/contributors).
- [**ctharvey**](https://github.com/ctharvey)
- [**dimasd-angga**](https://github.com/dimasd-angga) (Dimas D. Angga)
- [**Evan-Pycraft**](https://github.com/Evan-Pycraft)
- [**ForeverYoungPp**](https://github.com/ForeverYoungPp)
- [**GTC2080**](https://github.com/GTC2080) (TaoMu)
- [**kabaka9527**](https://github.com/kabaka9527)
- [**lisniuse**](https://github.com/lisniuse) (Richie)
- [**wade19990814-hue**](https://github.com/wade19990814-hue)
- [**wviana**](https://github.com/wviana) (Wesley Viana)
Also a separate thank-you to [**Bernardxu123**](https://github.com/Bernardxu123)
for designing the project logo, and to
[AIGC Link](https://xhslink.com/m/80ngts127cA) for promoting the project on XiaoHongShu.
