Official · AI/ML · 1 db · 1 docker
litellm
litellm
The AI Gateway for platform teams
preflight
topology
How this stack is wired
Each card is a service this deploy creates. A line means one service reads another through env.
listen
Ports the stack publishes
Service on the left. Host port, then the container port and scheme.
- litellm80→4000/http
you set
Values this template leaves for you
Literals and placeholders. Wiring to another service lives on the topology, not here.
STORE_MODEL_IN_DB
litellm
True
notes
Context for this stack
Written with the template. Topology, binds, and values you set live in the sections above.
LiteLLM is an open-source LLM gateway that provides a unified API for accessing 100+ AI models and providers using the OpenAI-compatible input and output format.
About Hosting LiteLLM
This dFlow template deploys LiteLLM with a PostgreSQL database for storing application data, configuration, usage, and spend information. Connect multiple AI providers and manage them through a single API and dashboard.
Common Use Cases
- Access 100+ LLMs and AI providers through one API
- Use a consistent OpenAI-compatible input and response format
- Route requests across OpenAI, Azure, Anthropic, Gemini, and other providers
- Configure retry and fallback logic across multiple deployments
- Track AI spending and usage
- Set budgets and manage costs per project
- Access endpoints for chat, responses, embeddings, images, audio, batches, and more
Dependencies
- LiteLLM
- PostgreSQL
Why Deploy on dFlow?
Deploy LiteLLM in minutes with automatic HTTPS, persistent storage, environment variable management, custom domains, and centralized application management—all from a single dashboard.