Skip to content

Local Quickstart

Use this guide to run TraceLLM locally and send a real LLM trace from the example user application.

Prerequisites

  • Node.js 20+
  • pnpm 9+
  • Docker Desktop
  • One real LLM provider key:
  • OpenAI, or
  • Anthropic Claude, or
  • Gemini

Install Dependencies

Run from the project root:

cd C:\dev\tracellm
pnpm install

Start TraceLLM Backend

pnpm infra:build
pnpm infra:up

Backend health:

http://localhost:4319/health

Swagger:

http://localhost:4319/api-docs

Start Web UI

In another terminal:

pnpm dev:web

Open the web URL printed by Vite. It is usually:

http://localhost:5173

Create Account

In the TraceLLM UI:

  1. Register a new account.
  2. Open Project Config.
  3. Create or copy an API key.
  4. Save the API key for the example app.

Run Real LLM Example

Set common TraceLLM variables:

$env:TRACELLM_ENDPOINT="http://localhost:4319"
$env:TRACELLM_API_KEY="trllm_your_key_from_the_ui"

Choose a provider.

$env:LLM_PROVIDER="openai"
$env:OPENAI_API_KEY="sk_your_openai_key"
$env:OPENAI_MODEL="gpt-4.1-mini"
pnpm example:user-app

$env:LLM_PROVIDER="anthropic"
$env:ANTHROPIC_API_KEY="sk-ant-your_key"
$env:ANTHROPIC_MODEL="claude-3-5-sonnet-latest"
pnpm example:user-app

$env:LLM_PROVIDER="gemini"
$env:GEMINI_API_KEY="your_gemini_key"
$env:GEMINI_MODEL="gemini-1.5-flash"
pnpm example:user-app

Confirm Success

Refresh the TraceLLM UI.

You should see:

Customer support answer workflow

The timeline should include:

  • fetch.customer.profile
  • fetch.customer.activity
  • provider LLM span such as openai.chat.generate
  • events and token usage when provider metadata is available