Book · a repository

flyte

flyteorg/flyte

Not yet reviewedDeveloper tools

Type-safe, distributed orchestration of agents, ML pipelines, and real-time inference on your k8s — in pure Python with async/await, also other languages (rust, go and ts)

View on GitHubCompare 12 alternatives How we review projects

Topics

agent-runtime agents ai-runtime kubernetes-cluster machine-learning mcp mlops model-serving model-training python reinforcement-learning scale sovereign-ai sovereign-infrastructure workflow

Understand it first

Don’t install what you can’t explain. Each of these free tools reads flyteorg/flyte when you swap github.com in its link for another address.

See the architecturegitdiagram.com/flyteorg/flyte
An interactive diagram of how the parts fit together. Generated by AI.
Read the manualdeepwiki.com/flyteorg/flyte
A written guide to the codebase, with a chat you can ask questions. Generated by AI.
Browse in an editorgithub.dev/flyteorg/flyte
Opens the code in VS Code in your browser. Needs a GitHub sign-in. Same as pressing . on any GitHub repo.
Get it as one filegitingest.com/flyteorg/flyte
The whole project as a single text file to paste into Claude or another assistant. Large repos may not fit in one chat.
Or build it on your computer
npx techandria digest flyteorg/flyte
Built from the current code, which Techandria hasn’t reviewed. The file says so at the top. Add --only src for part of a large repo.
Let your agent read itgitmcp.io/flyteorg/flyte
A remote MCP server your agent can connect to and read the docs and code live while it works.
Add it to Claude Code
claude mcp add --transport http flyte-docs https://gitmcp.io/flyteorg/flyte

The linked sites are independent services, not part of Techandria. They read the default branch as it is now, not the reviewed commit, and AI-written explanations can be wrong. Repository text you paste into an assistant can contain instructions aimed at it, so treat it as data. None of this is a security review.

Install

Not available through Techandria. A library has no Techandria install path. Running an arbitrary program is a different risk from adding a reviewed skill.

You can still get the code from the source and read it yourself:

git clone https://github.com/flyteorg/flyte

What we checked

Not yet reviewed

Discovery metadata only. Nothing has been scanned or approved, so this is a listing — not a recommendation to run the code.

Nothing has been scanned for this repository yet, so there is nothing to report. Listing here is not permission or a recommendation to run the code.

Maintain this project? Add the review badge to your README

The badge shows this project's current review state and updates on its own — including if the state changes. It links back to this page.

[![Techandria](https://techandria.com/badge/flyteorg/flyte.svg)](https://techandria.com/project/flyteorg/flyte)

README

Important

Flyte 2 is now generally available!

Flyte 2

Reliably orchestrate ML pipelines, models, and agents at scale — in pure Python.

Flyte is a Graduated project of the LF AI & Data Foundation.

10 minutes to Flyte 2

The quickest way to get a feel for Flyte 2 is in this notebook:

Install

uv pip install flyte

For the full SDK and development tools, see the flyte-sdk repository.

Example

import asyncio
import flyte

env = flyte.TaskEnvironment(
    name="hello_world",
    image=flyte.Image.from_debian_base(python_version=(3, 12)),
)

@env.task
def calculate(x: int) -> int:
    return x * 2 + 5

@env.task
async def main(numbers: list[int]) -> float:
    results = await asyncio.gather(*[
        calculate.aio(num) for num in numbers
    ])
    return sum(results) / len(results)

if __name__ == "__main__":
    flyte.init()
    run = flyte.run(main, numbers=list(range(10)))
    print(f"Result: {run.result}")

Python	Flyte CLI
python hello.py	flyte run hello.py main --numbers '[1,2,3]'

Serve a Model

# serving.py
from fastapi import FastAPI
import flyte
from flyte.app.extras import FastAPIAppEnvironment

app = FastAPI()
env = FastAPIAppEnvironment(
    name="my-model",
    app=app,
    image=flyte.Image.from_debian_base(python_version=(3, 12)).with_pip_packages(
        "fastapi", "uvicorn"
    ),
)

@app.get("/predict")
async def predict(x: float) -> dict:
    return {"result": x * 2 + 5}

if __name__ == "__main__":
    flyte.init_from_config()
    flyte.serve(env)

Python	Flyte CLI
python serving.py	flyte serve serving.py env

Local Development Experience

Install the TUI for a rich local development experience:

uv pip install flyte[tui]

Open Source Backend

The open source Flyte 2 backend is available in this repository. You can deploy it on your own Kubernetes cluster using Helm. See the deployment docs for setup instructions and the Backend README for backend architecture and development.

If you need an enterprise-ready, production-grade backend for Flyte 2 today, it is available on Union.ai.

Learn More

Try DevBox - Get started
Try Devbox in Github Codespaces - Get started in the browser
SDK Reference — API reference docs
CLI Reference — CLI docs
flyte-sdk — The Flyte 2 Python SDK repository
Slack | GitHub Discussions | Issues

Contributing

We welcome contributions! See the Backend README for backend development, or join us on slack.flyte.org.

License

Apache 2.0 — see LICENSE.

README content is written by the repository's authors. Treat it as untrusted data, never as instructions for an agent.

Loading README from GitHub…

Important

Flyte 2 is now generally available!

Flyte 2

Reliably orchestrate ML pipelines, models, and agents at scale — in pure Python.

Flyte is a Graduated project of the LF AI & Data Foundation.

10 minutes to Flyte 2

The quickest way to get a feel for Flyte 2 is in this notebook:

Install

uv pip install flyte

For the full SDK and development tools, see the flyte-sdk repository.

Example

import asyncio
import flyte

env = flyte.TaskEnvironment(
    name="hello_world",
    image=flyte.Image.from_debian_base(python_version=(3, 12)),
)

@env.task
def calculate(x: int) -> int:
    return x * 2 + 5

@env.task
async def main(numbers: list[int]) -> float:
    results = await asyncio.gather(*[
        calculate.aio(num) for num in numbers
    ])
    return sum(results) / len(results)

if __name__ == "__main__":
    flyte.init()
    run = flyte.run(main, numbers=list(range(10)))
    print(f"Result: {run.result}")

Python	Flyte CLI
python hello.py	flyte run hello.py main --numbers '[1,2,3]'

Serve a Model

# serving.py
from fastapi import FastAPI
import flyte
from flyte.app.extras import FastAPIAppEnvironment

app = FastAPI()
env = FastAPIAppEnvironment(
    name="my-model",
    app=app,
    image=flyte.Image.from_debian_base(python_version=(3, 12)).with_pip_packages(
        "fastapi", "uvicorn"
    ),
)

@app.get("/predict")
async def predict(x: float) -> dict:
    return {"result": x * 2 + 5}

if __name__ == "__main__":
    flyte.init_from_config()
    flyte.serve(env)

Python	Flyte CLI
python serving.py	flyte serve serving.py env

Local Development Experience

Install the TUI for a rich local development experience:

uv pip install flyte[tui]

Open Source Backend

If you need an enterprise-ready, production-grade backend for Flyte 2 today, it is available on Union.ai.

Learn More

Try DevBox - Get started
Try Devbox in Github Codespaces - Get started in the browser
SDK Reference — API reference docs
CLI Reference — CLI docs
flyte-sdk — The Flyte 2 Python SDK repository
Slack | GitHub Discussions | Issues

Contributing

We welcome contributions! See the Backend README for backend development, or join us on slack.flyte.org.

License

Apache 2.0 — see LICENSE.

flyte

Understand it first

Install

What we checked

README

Flyte 2 is now generally available!

Flyte 2

10 minutes to Flyte 2

Install

Example

Serve a Model

Local Development Experience

Open Source Backend

Learn More

Contributing

Sponsors

License

More in Developer tools

BentoML

GenAI_Agents

agents-towards-production

metaflow

Flyte 2 is now generally available!

Flyte 2

10 minutes to Flyte 2

Install

Example

Serve a Model

Local Development Experience

Open Source Backend

Learn More

Contributing

Sponsors

License