Flowcat is an Apache-2.0 native-Rust runtime for real-time voice AI agents (phone + WebRTC). It runs a full voice pipeline — STT · LLM · TTS, or a single speech-to-speech model — as one self-contained binary you host in your own VPC or fully air-gapped. No hosted control plane, no per-minute billing; bring your own provider keys and a call's audio never leaves your infra. pipecat-compatible pipeline, native SIP/RTP, and dense enough to hold thousands of concurrent calls per box.
No reviews yetBe the first to leave a review for Flowcat
Maker
📌
Hi PH! I built Flowcat because most voice-agent platforms are hosted SaaS — your call audio flows through someone else's cloud and you pay per minute. Flowcat is the opposite: one Rust binary you own and run yourself. It's a clean-room counterpart to pipecat's architecture, compiled to a single static binary — no GC/GIL, so one process stays flat across thousands of concurrent calls. Apache-2.0, pre-1.0, built in the open. Writeup + benchmark: https://dev.to/iam_suriyan_b9078... — would love your feedback!
Report
Spun up the binary locally with a Pipecat pipeline and was honestly surprised how little config it took to get SIP calls flowing. The fully self-contained, no-control-plane angle is a nice change of pace from the usual cloud-tied stacks.
Spun up the binary locally with a Pipecat pipeline and was honestly surprised how little config it took to get SIP calls flowing. The fully self-contained, no-control-plane angle is a nice change of pace from the usual cloud-tied stacks.