This document covers production deployment strategies for AutoGPT Platform, including Docker orchestration, service dependencies, and environment management. The deployment architecture is designed for scalable, production-ready operation with proper service isolation and monitoring.
For development setup and initial configuration, see Getting Started. For infrastructure-level CI/CD processes, see CI/CD Workflows.
AutoGPT Platform uses a microservices architecture deployed via Docker Compose, with clear separation between application services, data services, and external dependencies.
Sources: autogpt_platform/docker-compose.platform.yml51-277 autogpt_platform/docker-compose.yml26-169
The system relies on specific environment variables to define service discovery within the Docker network.
Sources: autogpt_platform/docker-compose.platform.yml9-35 autogpt_platform/docker-compose.platform.yml100-137
The deployment uses strict service dependencies with health checks to ensure proper startup ordering.
| Service | Depends On | Health Check Mechanism |
|---|---|---|
migrate | db | prisma migrate status autogpt_platform/docker-compose.platform.yml77-82 |
redis-init | redis-0, 1, 2 | redis-cli cluster info autogpt_platform/docker-compose.platform.yml131-137 |
db | rabbitmq | rabbitmq-diagnostics ping autogpt_platform/docker-compose.yml166-168 |
rest_server | db, redis-0, rabbitmq | Container Health Status |
For a deep dive into the Compose configuration, see Docker Compose Deployment.
Configuration is managed via .env files, with backend/.env.default providing the base values.
For a full reference of available variables, see Environment Configuration.
Sources: autogpt_platform/backend/.env.default7-38 autogpt_platform/docker-compose.platform.yml18-35
The backend autogpt_platform/backend/Dockerfile uses a sophisticated multi-stage build to minimize image size and provide a secure runtime.
builder: Installs build dependencies (Python 3.13, Node.js, Poetry) and generates the Prisma client autogpt_platform/backend/Dockerfile3-54migrate: A lightweight stage containing only the Prisma CLI and migration files for database updates autogpt_platform/backend/Dockerfile59-86server: The final production image. It includes runtime dependencies like ffmpeg, imagemagick, and bubblewrap for secure tool execution autogpt_platform/backend/Dockerfile90-109The frontend autogpt_platform/frontend/Dockerfile uses a standard Next.js standalone build:
base: pnpm dependency installation autogpt_platform/frontend/Dockerfile2-6build: Generates the API client and runs next build autogpt_platform/frontend/Dockerfile9-33prod: Minimal runtime using the standalone output autogpt_platform/frontend/Dockerfile36-56The platform uses Prisma for schema management and migrations. Production deployments must run the migrate service to ensure the schema is up to date before application services start.
Key Operations:
migrate service using prisma migrate deploy autogpt_platform/docker-compose.platform.yml52-62python3 scripts/gen_prisma_types_stub.py is required for backend type safety autogpt_platform/backend/Dockerfile54For detailed database procedures, see Database Management.
bubblewrap to provide OS-level sandboxing for the bash_exec tool when external sandboxes like E2B are not configured autogpt_platform/backend/Dockerfile108The repository includes GitHub Action workflows for automated testing and validation of the deployment artifacts:
Sources: autogpt_platform/README.md141-165 autogpt_platform/backend/Dockerfile96-109
Refresh this wiki