We are open-sourcing Wavefront AI, the AI middleware built over FloAI.

We have been building flo-ai for more than an year now. We started the project when we wanted to experiment with different architectures for multi-agent workflows.

We started with building over Langchain, and eventually realised we are getting stuck with lot of langchain internals, for which we had to do a lot of workrounds. This forced us to move out of Langchain & and build something scratch-up, and we named it flo-ai. (Some of you might have already seen my previous posts on flo-ai)

We have been building production use-case using flo-ai for last year, and taking the same to production. At this point the agents where performing well, but the next problem was to connect agents to different data sources and service available in enterprises, thats when we built wavefront.

Wavefront is an AI middleware platform designed to seamlessly integrate AI-driven agents, workflows, and data sources across enterprise environments. It acts as a connective layer that bridges modular frontend applications with complex backend data pipelines, ensuring secure access, observability, and compatibility with modern AI and data infrastructures.

We are now open-sourcing wavefront, and its coming in the same repository as flo-ai.

We have just updated the README for the same, showcasing the architecture and a glimpse of whats about to come.

We are looking for feedback & some early adopters.

Please join our discord(https://discord.gg/BPXsNwfuRU) to get latest updates, share feedback and to have deeper discussions on use-cases.

Release: Dec 2025
Give us a star @ https://github.com/rootflo/wavefront