Astral is an open-source framework for AI engineers that abstracts away the complexity and friction of working across multiple model providers.
Explore the docs »
View Demo
·
Home Page
·
Request Feature
Table of Contents
Astral is an open-source framework for AI engineers that abstracts away the complexity and friction of working across multiple model providers.
Astral comes from Late Latin, meaning of, relating to, or coming from the stars. The stars have long symbolized discovery, exploration, and the interconnected systems that shape our universe.
AI capabilities such as completions, embeddings, image and video generation, real-time audio, and speech-to-text are essential building blocks of the next industrial revolution. But today, these capabilities remain fragmented, locked behind provider-specific implementations that slow innovation.
Astral changes that. By seamlessly integrating these powerful resources into a cohesive, universal framework, Astral enables engineers to build at the speed of discovery. We empower innovators to transcend today's limitations, guiding them toward a new frontier of human-AI interaction—unlocking infinite possibilities.
Astral provides a type-safe, unified interface that developers can use to integrate across providers and resources, including completions, embeddings, real-time audio, and speech-to-text, without being locked into any single provider. By eliminating provider-specific inconsistencies, Astral enables engineers to build, scale, and iterate on AI-driven applications with greater speed and efficiency.
Looking ahead, we envision building enterprise solutions on top of Astral's core open-source SDK, including prompt versioning, template management, workflow automation, and evaluation frameworks that work across providers and organizations.
Our goal is to become the go-to framework for building AI applications, from agentic systems to enterprise workflows, while remaining a thin, high-performance layer that avoids the inefficiencies of heavier frameworks.
This section shows you how to get a local copy up and running. For a more structured overview, see our docs here: Astral Documentation
- Requires Python >=3.12
- Authentication credentials for whatever model providers you're interested in using (i.e API key for OpenAI, API key for Anthropic, etc)
Before diving in, make sure to:
-
⭐️ Star the Repository!
Show your support by starring the repo—it helps new builders discover us and expands our community! -
💬 Join our Discord Community:
Hop into our vibrant Discord community to connect with fellow contributors, brainstorm ideas, ask questions, or just hang out! It's the best place to stay updated on everything Astral. -
🐦 Follow me on X:
Stay updated with the latest news and updates by following me on X.Thanks for helping shape the future of Astral—we can't wait to see what you build!
Ready to build something awesome? Follow these steps:
-
🍴 Fork and Clone:
Begin by forking the repository to your GitHub account and cloning your fork locally:# First, fork the repository on GitHub, then clone your fork git clone https://github.com/chrismaresca/Astral-AI.git cd Astral-AI
-
⚙️ Set Up Development Environment:
We utilize UV—a high-performance Python package and project manager built with Rust—to simplify setup.# Execute the following command to create a virtual environment and install all dependencies. uv sync -
📚 Refer to Quick Start Guide:
Access our developer quick start guide here: Quick Start
Dive into our demos and guides to get a hands-on experience: Demos
See our vision for the future of AI engineering: Future Vision 🌌
See the open issues for a full list of proposed features (and known issues) 📝.
If you found a bug or have a feature request? Check out our Found an issue? section to learn how to contribute.
Your contributions are what make Astral shine! For more information, see the CONTRIBUTING.md file.
If you have found an issue with our documentation, please create an issue.
If it's a quick fix, such as a misspelled word or a broken link, feel free to skip creating an issue. Go ahead and create a pull request with the solution. 🚀
Or you can contact Chris directly if you prefer.
Distributed under the MIT License. See the LICENSE file for more information.
- Building Effective Agents – Anthropic - Insights on creating autonomous AI systems that can effectively solve complex tasks
- Aligning Language Models To Follow Instructions – OpenAI - Best practices for crafting effective prompts
- Linear – In artfully blending software and design - Inspiration for thoughtful product design and user experience
- Steve Jobs on Changing the World - "When you grow up you tend to get told the world is the way it is... life changes when you realize one simple truth: Everything around you that you call life was made up by people that were no smarter than you."
- My approach to software development has been profoundly shaped by the philosophical and poetic insights of Charles Bukowski:
- Style - Charles Bukowski - "Style is the answer to everything..."
- So You Want To Be a Writer - Charles Bukowski - On authentic creation and finding your voice