LLM Observer Documentation

Installation

macOS (Homebrew)

The fastest way to install LLM Observer on macOS:

brew install llm-observer-dev/tap/llm-observer

Linux

Download the latest release for your architecture:

# AMD64
curl -L https://github.com/llm-observer-dev/llm-observer/releases/latest/download/llm-observer_linux_amd64.tar.gz | tar xz
sudo mv llm-observer /usr/local/bin/

# ARM64
curl -L https://github.com/llm-observer-dev/llm-observer/releases/latest/download/llm-observer_linux_arm64.tar.gz | tar xz
sudo mv llm-observer /usr/local/bin/

Windows

Download the latest .exe from the releases page and add it to your PATH. Or use scoop:

scoop install llm-observer

First Run

After installing, run:

llm-observer

LLM Observer will automatically detect your local AI tool history (Claude Code, Cursor, Aider, etc.) and launch the dashboard at http://localhost:8089.

Configuration

LLM Observer stores configuration at ~/.llm-observer/config.json.

{
  "port": 8089,
  "host": "127.0.0.1",
  "providers": "auto",
  "notifications": {
    "enabled": true,
    "budgetAlerts": true
  }
}

Provider Overview

LLM Observer supports tracking costs across multiple tools by reading their local data files.

• Claude Code: Reads JSONL session history.
• Cursor IDE: Reads local SQLite tracking database.
• Aider: Reads structured analytics logs.

Sessions Tab

Browse every AI conversation with full metadata. Sort by cost to find expensive sessions instantly. Sortable columns include Date, Provider, Model, Duration, and Estimated Cost.

KPI Cards

Visualize your LLM usage through high-level metrics including total tokens, estimated cost, and average latency per provider.

Costs Tab

Detailed breakdown of expenditures across specific timeframes. Filter by project or model to optimize your AI spend.