Agentic Workflows Can Use Local Inference
Agentic Workflows gives you complete control to customize your Actions runtime and choose your AI model. You can deploy Agentic Workflows to a custom runner of your choosing and use any compatible AI host or model-routing platform. You can even use local inference running on your Mac!
Because Agentic Workflows compile to normal Actions jobs, they can run wherever Actions runs, including on your own custom runners on any host of your choice. And just like you can choose your own agentic harness, it’s also easy to point them to a different model or model-routing platform.
If you’re one of the growing cohort who run their own AI models, you can even run your Agentic Workflows directly on your Mac using your own local inference.
We’ve put up a couple demonstration workflows to help you get started. Let us know where you’re running your Agentic Workflows on the GitHub Next Discord.