labubu 0.1.0

Local-first LLM observability platform

Labubu

Local-first LLM observability platform.

Installation

From PyPI (recommended)

pip install labubu

PyPI hosts one wheel per platform (the wheel bundles the Go binary), so pip auto-selects the correct one.

From GitHub Release

Install the wheel matching your platform directly from GitHub Releases:

# Linux x86_64
pip install https://github.com/Wendymayu/labubu/releases/download/v0.1.0/labubu-0.1.0-py3-none-linux_x86_64.whl

# Linux ARM64
pip install https://github.com/Wendymayu/labubu/releases/download/v0.1.0/labubu-0.1.0-py3-none-linux_aarch64.whl

# macOS Intel
pip install https://github.com/Wendymayu/labubu/releases/download/v0.1.0/labubu-0.1.0-py3-none-macosx_10_9_x86_64.whl

# macOS Apple Silicon
pip install https://github.com/Wendymayu/labubu/releases/download/v0.1.0/labubu-0.1.0-py3-none-macosx_11_0_arm64.whl

# Windows
pip install https://github.com/Wendymayu/labubu/releases/download/v0.1.0/labubu-0.1.0-py3-none-win_amd64.whl

Replace v0.1.0 with the desired version. Available releases: github.com/Wendymayu/labubu/releases

Standalone Binary

Download the raw binary for your platform from the Releases page:

Platform	File
Linux x86_64	labubu-linux-amd64
Linux ARM64	labubu-linux-arm64
macOS Intel	labubu-darwin-amd64
macOS Apple Silicon	labubu-darwin-arm64
Windows	labubu-windows-amd64.exe

Quick Start

labubu serve

Then open http://localhost:8080 for the UI.

Configure your agent to send OTLP traces to http://localhost:4318.

Commands

Command	Description
labubu serve	Start the server (default port 8080)
labubu version	Print version information
labubu help	Show help

Serve Options

Flag	Default	Description
--port	8080	API & UI listen port
--log-level	info	Log level: debug, info, warn, error
--buffer-size	1000	Pipeline buffer capacity
--flush-interval	200ms	Pipeline flush interval

Example

labubu serve --port 9090 --log-level debug