Skip to content

Launcher Wizard Guide

Your DevOps steward/conscierge that spins up complete Atlas applications in minutes.

Overview

The Launcher Wizard is your DevOps steward/conscierge that spins up complete Atlas applications in minutes.

What It Does:

  1. Spins up infrastructure — Kubernetes namespaces with your own security and sandbox
  2. Deploys microservices - Bot, AG-UI Gateway, Engine, Elasticsearch, Redis, etc.
  3. Exposes URLs - Links to engine, components, admin UI
  4. Provides env vars - Download configuration for local development

Result: A fully functional Atlas application running in your own Kubernetes namespace, ready to use in minutes.

Workflow

  1. Go to launcher - https://atlas.gurunetwork.ai/launcher
  2. Follow wizard - 4 steps (Identity, Bot Setup, Group Setup, Ready)
  3. Get your app - Running in k8s namespace, URLs exposed
  4. Fork the repo - Switch infrastructure to your own fork
  5. Develop locally - Use env vars from /admin command in bot

The 4 Steps

Step 1: Community Identity (25%) - Define bot name, username, description, and type (Community or Personal)

Step 2: Connect Your Telegram Bot (50%) - Get bot token from @BotFather (/newbot) - Paste token into launcher

Step 3: Group Setup (75%) - Add bot to your Telegram group - Grant necessary permissions - Deployment starts automatically

Step 4: Ready (100%) - Deployment complete - Access bot, admin UI, and get env vars for local development

What Gets Deployed

Stateless Microservices

  • bot-app — Telegram bot + AG-UI Gateway
  • webchatbot-app — AI assistant for web
  • engine-api — BPMN workflow engine
  • flowapi-api — Auth, app config, analytics API
  • warehouse-api — WebSocket event stream

Stateful Infrastructure

  • PostgreSQL (via PgBouncer) — Primary database
  • Redis — Cache and state management
  • Elasticsearch — Knowledge bases

Optional (Enterprise)

  • ClickHouse — Event bus and analytics
  • RabbitMQ — Async workloads
  • Workers — External task workers

Deployment: Each application runs in its own Kubernetes namespace (atlas-{app-id}) with isolated resources, security policies, and monitoring.

Exposed URLs

After deployment, the launcher UI displays:

  • Bot URL — https://t.me/{bot_username}
  • Engine API — https://engine-{app-id}.atlas.gurunetwork.ai
  • FlowAPI — https://flowapi-{app-id}.atlas.gurunetwork.ai
  • Warehouse API — wss://warehouse-{app-id}.atlas.gurunetwork.ai
  • Admin UI — https://admin-{app-id}.atlas.gurunetwork.ai
  • Webchatbot — Embeddable widget URL

Development Environment Setup

Get Environment Variables

  1. In Telegram Bot:
  2. Send /admin command

  3. Bot responds with download link for env vars

  4. Download Configuration:

  5. Click link to download .env file

  6. Contains all connection strings and secrets

  7. Use for Local Development: ```bash # Copy env vars cp .env.local .env

# Run development script ./run_development.sh ```

What's Included: - Database URLs (PostgreSQL, Redis, Elasticsearch) - Service URLs (FlowAPI, EngineAPI, warehouse-api) - API keys and tokens - Feature flags

Fork and Customize

Step 1: Fork the Repository

# Fork on GitHub
https://github.com/evahteev/sol-atlas

# Clone your fork
git clone https://github.com/your-username/sol-atlas
cd sol-atlas

Step 2: Switch Infrastructure

  1. Update Kubernetes Configs:
  2. Point to your own cluster
  3. Update namespace references

  4. Configure your own ingress

  5. Update Environment Variables:

  6. Use your own database URLs
  7. Configure your own service URLs

  8. Set your own secrets

  9. Deploy: bash kubectl apply -f k8s/

Step 3: Keep Developing

  • Workflows — Edit BPMN diagrams or YAML
  • Personas — Edit Context.md + YAML
  • Tools — Add LangGraph tools or custom code
  • UI — Customize webchatbot, admin portal

Benefits: - Full control over infrastructure - Custom deployments - Keep using launcher for new apps - Develop against your own fork

Benefits

  • Minutes to production - Not weeks or months
  • Your own namespace - Isolated, secure, sandboxed
  • Portable - Fork repo, switch infra, keep developing
  • Templates - Pre-built use cases (community bot, personal bot, intercom)

Use Cases

  • Quick start - Spin up bot, test, iterate
  • Development environment - Local dev against deployed infra
  • Production deployment - One-click deploy to your k8s cluster
  • Template customization - Start from template, customize workflows

Templates

Community Bot — Quest system, leaderboards, wallet integration, analytics dashboard

Personal Bot — Daily briefs, custom routines, personal notes, knowledge base

Intercom Bot — Webchatbot widget, Telegram support group integration, knowledge base RAG, escalation workflows

Troubleshooting

Deployment Fails

  • Check Kubernetes cluster access and resource availability
  • Review launcher logs and Kubernetes events

Bot Not Responding

  • Verify bot token in BotFather
  • Re-add bot to group with necessary permissions
  • Check bot logs in admin UI

Can't Access Admin UI

  • Verify URL in launcher UI
  • Check network connectivity and browser console

Best Practices

Security: - Never share bot tokens or API keys - Use secrets for sensitive data - Enable RBAC in Kubernetes

Performance: - Monitor resources (CPU, memory) - Scale horizontally for stateless services - Optimize databases for stateful services

Development: - Version control workflows and configs - Test locally before deploying - Monitor logs for debugging

Next Steps

Support

  • Documentation: https://atlas.gurunetwork.ai/docs
  • Community: https://t.me/SolanaAtlas
  • Issues: https://github.com/evahteev/sol-atlas/issues