Build apps

An LLM-optimized bundle of this entire section is available at section.md. This single file contains all pages in this section, optimized for AI coding agent context.

This section covers how to build different types of apps with Flyte, from single-script apps to multi-file projects, common usage patterns, and authentication.

Go to Introducing apps for an overview of apps and a quick example. For pre-built environments for popular frameworks like Streamlit, FastAPI, vLLM, and SGLang, see Native app integrations.

App types

Flyte supports various types of apps:

  • UI dashboard apps: Interactive web dashboards and data visualization tools like Streamlit and Gradio
  • Web API apps: REST APIs, webhooks, and backend services like FastAPI and Flask
  • Model serving apps: High-performance LLM serving with vLLM and SGLang

For ready-to-use environments for these frameworks, see Native app integrations.

Usage patterns

Apps and tasks can interact in various ways: calling each other via HTTP, webhooks, WebSockets, or direct browser usage.

Pattern Use Case Implementation
App Stand-alone serving app HTTP requests from arbitrary clients
App → App Microservices, proxies, agent routers, LLM routers HTTP requests between apps
App → Task Webhooks, APIs triggering workflows Flyte SDK in app
Task → App Batch processing using inference services HTTP requests from task
Browser app User-facing dashboards (e.g. Streamlit, Gradio) Direct browser access

Next steps