VoiceBlender is an open source framework designed to bridge SIP and WebRTC voice calls, providing a toolkit for building voice-driven applications with built-in AI integration. Developed as a Go service, it enables multi-party audio mixing and supports both inbound and outbound SIP calls, as well as browser-based WebRTC connections without the need for plugins. The platform is intended for developers and organizations looking to create solutions such as IVR systems, contact centers, and multilingual help lines, leveraging real-time voice transformation and AI-powered capabilities.
The framework offers a REST API and real-time webhooks for programmatic control, allowing users to originate and receive SIP calls, manage multi-party rooms, and connect browsers via SDP offer/answer with trickle ICE. VoiceBlender supports multiple audio codecs, including PCMA, PCMU, and Opus. Users can plug in their own TTS (text-to-speech), STT (speech-to-text), and AI agent models, with built-in support for providers such as ElevenLabs, Google Cloud, AWS Polly, Deepgram, VAPI, and Pipecat. The platform enables real-time STT with partial transcripts and allows conversational AI agents to be attached to individual call legs or entire rooms, supporting features like barge-in.
Additional capabilities include integration with WhatsApp Business for inbound and outbound voice calls, support for DTMF input, audio playback, call recording, and multilingual voice hotlines. The API provides endpoints for managing calls (legs), rooms, and agent interactions, as well as for sending messages to AI agents. VoiceBlender also facilitates the configuration of environment variables for SIP, ICE servers, webhooks, and provider API keys, with options for local or Docker-based deployment. Performance benchmarks indicate low audio latency and high concurrent room setup throughput.
VoiceBlender is distributed as free and open source software, with its source code and documentation available on GitHub. Its open design and modular architecture make it suitable for a wide range of voice communication and automation scenarios.
VoiceBlender is a Voice, TTS & speech product. It focuses on integrating SIP, WebRTC, and AI voice features is complex for developers building communication platforms. VoiceBlender is an open-source project aimed at voice application developers. The project is open source (MIT). The product ships for the web, API, and the command line, and it can be self-hosted.
VoiceBlender first shipped in 2026. The project is developed in the open on GitHub with 83 stars and 88 commits in the last 90 days. Across PulseGate's embedding index, VoiceBlender has few near neighbours, marking it as relatively distinct. Among its 8 catalogued features are SIP voice calls, webRTC support, and TTS and STT integration. It exposes integrations via a public API.
Latest indexed changes and source events
voiceblender.org discovered by the PulseGate indexer
Other apps tracked under the same category.
