Pipecat.
A vendor-neutral framework for realtime voice and multimodal agents with broad STT, LLM, TTS, transport, and telephony integrations.
What you are adopting
Repository and documentation checked 2026-07-17. Always re-check license and release activity before adoption.
When Pipecat belongs on the shortlist
The strongest framework is the one that matches your ownership boundary, not the one with the longest provider list.
Best fit
Provider-neutral orchestration and teams that want to own the full media pipeline.
Operational trade-off
Production hosting, scaling, observability, and carrier operations remain your responsibility unless you use Pipecat Cloud.
Cost model
Open-source code does not make the full call free. Budget separately for speech models, language models, telephony, phone numbers, recording, storage, observability, compute, support, and the engineering time needed to keep a realtime service healthy.
Inspect the project directly
Repository metadata can drift. Use the source and current documentation for a final technical or legal decision.