Jump to content

Connect Leeroopedia MCP: Equip your AI agents to search best practices, build plans, verify code, diagnose failures, and look up hyperparameter defaults.

Implementation:Helicone Helicone Dev Server Startup

From Leeroopedia
Knowledge Sources
Domains Local Development, Services
Last Updated 2026-02-14 00:00 GMT

Overview

Concrete tool for starting the three main Helicone application services (Jawn, Web, Worker) in development mode, provided by Yarn scripts and Wrangler CLI.

Description

Each application service has its own development start command defined in its package.json:

  • Jawn (valhalla/jawn/package.json line 7): yarn dev runs concurrently "nodemon" "nodemon -x python3 genTypes.py" which starts the Express.js server with auto-reload and a parallel Python process for generating TypeScript types.
  • Web (web/package.json line 9): yarn dev:local runs next dev --turbo -p 3000 which starts the Next.js development server with Turbopack for fast rebuilds. An alternative yarn dev:better-auth (line 11) starts on port 3008 with Better Auth environment variables.
  • Worker (worker/wrangler.toml): Started via npx wrangler dev with --var WORKER_TYPE:OPENAI_PROXY flag. The default development configuration listens on port 8787 and connects to local Supabase, ClickHouse, MinIO, and Jawn.

Usage

Start services in order: Jawn first (port 8585), then Web (port 3000), and optionally Worker (port 8787). Each runs in its own terminal.

Code Reference

Source Location

  • Repository: Helicone
  • File: valhalla/jawn/package.json (lines 4-13), web/package.json (lines 8-20), worker/wrangler.toml (lines 1-107)

Signature

# Jawn backend
cd valhalla/jawn && yarn dev

# Web frontend
cd web && yarn dev:local

# Worker proxy
cd worker && npx wrangler dev --local --var WORKER_TYPE:OPENAI_PROXY --port 8787

Import

# Install all monorepo dependencies from the repository root
yarn install

# Wrangler is a dev dependency of the worker package
# No additional global installs needed

I/O Contract

Inputs

Name Type Required Description
Infrastructure services services Yes PostgreSQL, ClickHouse, MinIO must be running
.env files files Yes Environment variables for each service
node_modules directory Yes Dependencies installed via yarn install
WORKER_TYPE string Worker only Worker variant: OPENAI_PROXY, ANTHROPIC_PROXY, HELICONE_API, GATEWAY_API, GENERATE_API

Outputs

Name Type Description
Jawn API HTTP server Express.js backend at http://localhost:8585
Web Dashboard HTTP server Next.js frontend at http://localhost:3000 (or 3008 with better-auth)
Worker Proxy HTTP server Cloudflare Workers proxy at http://localhost:8787

Usage Examples

Basic Usage

# Terminal 1: Start the Jawn backend API
cd valhalla/jawn
yarn dev
# Output: Server listening on port 8585

Start Web Frontend

# Terminal 2: Start the Next.js web dashboard
cd web
yarn dev:local
# Output: Next.js dev server on http://localhost:3000

Start Web with Better Auth

# Alternative: Start web with Better Auth integration
cd web
yarn dev:better-auth
# Output: Next.js dev server on http://localhost:3008

Start Worker Proxy

# Terminal 3: Start the OpenAI proxy worker
cd worker
npx wrangler dev --local --var WORKER_TYPE:OPENAI_PROXY --port 8787
# Output: Worker listening on http://localhost:8787

Start Different Worker Types

# Anthropic proxy
npx wrangler dev --local --var WORKER_TYPE:ANTHROPIC_PROXY --port 8789

# Helicone API worker
npx wrangler dev --local --var WORKER_TYPE:HELICONE_API --port 8788

# Gateway API worker
npx wrangler dev --local --var WORKER_TYPE:GATEWAY_API --port 8790

Jawn Dev Script Breakdown

The Jawn yarn dev script runs two processes in parallel via concurrently:

concurrently "nodemon" "nodemon -x python3 genTypes.py"
  • nodemon: Watches TypeScript source files and restarts the Express.js server on changes.
  • nodemon -x python3 genTypes.py: Watches for changes and regenerates TypeScript types from the API schema.

Related Pages

Implements Principle

Requires Environment

Page Connections

Double-click a node to navigate. Hold to expand connections.
Principle
Implementation
Heuristic
Environment