Show HN: Simulate AI Chatbot Conversations

2 hours ago 1

Open source AI agent testing and conversation simulation platform

OneRun helps you test, evaluate, and improve your AI agents through realistic conversation simulation at scale.

  • Simulate realistic user conversations with your AI agents across diverse personas and scenarios
  • Generate evaluation datasets with judge-labeled conversations for testing and benchmarking
  • Create training data for fine-tuning with preference pairs, critique-and-revise triples, and clean JSONL exports
  • Automate QA testing by running hundreds of conversations per build to catch issues before production
  • Surface edge cases that manual testing misses through adversarial and varied interaction patterns
  • Docker and Docker Compose
  • Node.js 22+ (for local development)
  • Python 3.12+ (for local development)

Copy the environment variables:

# In the docker directory cd docker cp .env.example .env

Edit .env with your configuration values.

2. Start Temporal Services

First, start the Temporal server (required for workflow processing):

cd docker/temporal docker compose up -d

Verify Temporal is running:

From the main docker directory:

cd docker docker compose up

This will start:

  • Database (PostgreSQL on port 5432)
  • API (FastAPI on port 3001)
  • App (Next.js on port 3000)

4. Access the Application

cd api # Install dependencies uv sync # Run with auto-reload make dev
cd app # Install dependencies pnpm install # Run development server pnpm dev
cd api # Run migrations make migrate-up # Rollback migrations make migrate-down
├── api/ # FastAPI backend ├── app/ # Next.js frontend ├── docker/ # Docker Compose configurations │ ├── temporal/ # Temporal server setup │ └── docker-compose.yml ├── python-sdk/ # Python SDK for workflows └── README.md
# Stop OneRun services cd docker docker compose down # Stop Temporal services cd docker/temporal docker compose down
Read Entire Article