Defang is a deployment platform that lets developers deploy AI applications and cloud services to their own AWS, GCP, or Azure accounts using Docker Compose and a single command. It is designed for individual developers, teams, and agencies who want to ship AI agents, web apps, and APIs without manually managing cloud infrastructure.
How It Works
- Users define their application stack in a standard Docker Compose file, optionally adding Defang annotations for managed services such as Postgres, Redis, MongoDB, or LLM access through Bedrock, Vertex AI, or Azure AI.
- Running defang up deploys the entire stack to the user specified cloud account. The CLI automatically provisions IAM roles, a private VPC, load balancers, TLS certificates, encrypted secrets, and observability logging.
- The same Compose file works across AWS, GCP, and Azure without modification, enabling deployment to the developer own account, a client account, or many customer accounts with identical workflows.
Key Capabilities
- AI infrastructure integration: Native support for Amazon Bedrock, Google Vertex AI, and Azure AI services configured automatically through Compose file annotations with no manual setup required.
- Managed data services: Automated provisioning of managed Postgres with pgvector extension, Redis, and MongoDB instances as part of every deployment.
- Security by design: Each service gets its own least-privilege IAM role. Secrets are encrypted in the cloud native secret manager. Deployments run in private VPCs with locked-down security groups. Defang holds SOC 2 certification.
- Open source CLI and providers: The CLI and cloud provider implementations are open source, allowing users to inspect every command and verify exactly what runs in their cloud account.
- Framework agnostic: Works with any AI framework including CrewAI, LangGraph, AutoGen, and Strands, as well as standard web frameworks like Django, FastAPI, and Next.js.
Use Cases
- Individual developers deploying personal AI projects or side applications to their own cloud account with no DevOps overhead.
- Freelancers and agencies deploying applications into client cloud accounts and handing off projects with infrastructure included, avoiding per-client infrastructure work.
- Companies serving data-conscious customers who require AI agents to run inside each customer cloud account to maintain data sovereignty and compliance.
Pricing
- Hobby (free): Deploy to Defang Playground environment with limited scale and ephemeral storage. Suitable for hackathons and experimentation.
- Personal ($10 per user per month): Deploy to one personal cloud account with managed databases, custom domains, and no data reuse for AI training.
- Pro ($30 per user per month): Deploy to two clouds with GPU support, autoscaling, on-demand instances, and priority support.
- Enterprise (coming soon): Collaboration and control features for larger organizations.
