✻open source · MIT · v0.1

Your local LLM,
straight from your terminal.

A polished TUI chat client for self-hosted models. Streaming responses, multi-turn conversations, slash commands. Nothing leaves your machine.

Get started in 30s → Try it online ↓ GitHub
↳ works with Ollama ·llama.cpp ·LM Studio ·any OpenAI-compatible API
~/projects/chat — mini-claude
✻ Welcome to mini-claude


Built by
Hugo Stawiarski

model    llama3.2:3b
site     miniclaude.fr

Commands
/clear    start a fresh conversation
/install  install mini-claude locally


› 

enter send  ·  /clear reset  ·  /install install

↑ live demo powered by Claude — 3 messages on the web, unlimited locally.

Why mini-claude

Built for the terminal, by terminal people.

We don't try to do everything. We do one thing well: a flawless local chat experience.

🔒
Local-first by design
Zero round-trips to the cloud. Your conversations stay on your machine, forever. No telemetry. No accounts.
⚡
Token-by-token streaming
Real-time output via Server-Sent Events. The UI never blocks. Goroutines and channels, the Go way.
🔌
Works with anything
Targets the OpenAI-compatible HTTP API. Ollama, llama.cpp, LM Studio, vLLM — all plug-and-play.
⌨️
Slash commands
/model picks from your installed models, /clear resets the thread, /quit exits. Like Claude Code.
🎨
Polished by default
Built on Charm’s Bubble Tea + Lip Gloss. Smooth resize, spinner, multi-line input, scrollable history.
📦
One binary, zero deps
A single static Go binary. Or run it with Docker Compose alongside Ollama. Up in under a minute.

Install

Up and running in 30 seconds.

Make sure Ollama is running locally, pull a small model, then launch mini-claude.

ollama pull llama3.2:3b
git clone https://github.com/hugostarte/Mini-Claude.git
cd Mini-Claude && go run ./cmd/tui

Prefer Docker? docker compose run tui works too.

How it works

Three pieces. One goal.

01 — UI
Bubble Tea TUI

The model/view/update loop renders your chat. Textarea, viewport, spinner, slash command picker.

02 — Streaming
SSE over HTTP

A goroutine reads data: chunks from /v1/chat/completions and pipes tokens through a channel.

03 — Inference
Ollama (or any OpenAI API)

Runs locally. mini-claude only speaks HTTP. Swap the backend, swap the model — same client.

Built by Hugo Stawiarski

Open source · MIT · github.com/hugostarte/Mini-Claude