What is Dograh?
Dograh is an open-source, self-hostable platform for building voice AI agents. It serves as a transparent and customizable alternative to proprietary solutions like Vapi and Retell. Users can bring their own models, run everything on private infrastructure or use the managed cloud, and fully control the voice pipeline.
Core Features
The platform features a Model Context Protocol (MCP) server, allowing AI coding agents like Claude Code, Cursor, or OpenClaw to create, modify, and deploy complete voice agents directly from the IDE. It supports both traditional cascade (STT → LLM → TTS) and modern speech-to-speech pipelines using models like Gemini 3.1 Flash Live or GPT Realtime. The latter delivers ultra-low latency, natural turn-taking, and interruption handling.
A standout capability is blending pre-recorded human voice clips with TTS in the same cloned voice. The LLM intelligently chooses the best option, resulting in up to 2× higher conversion rates and up to 3× lower costs while sounding remarkably human.
Self-Hosted & Privacy-First
Full self-hosting support gives complete data sovereignty. You can run any compatible models (Whisper, Kokoro, Voxtral, etc.) inside your own VPC. This is one of the most requested features and a primary reason many teams choose Dograh.
Use Cases
Dograh is perfect for sales agents, EMI reminders, collections, customer support, appointment booking, and any scenario requiring natural, scalable voice conversations. The no-code workflow builder combined with deep programmability makes it suitable for both technical and non-technical teams.
Being fully open source, Dograh can be forked, extended, and deployed according to specific organizational requirements while maintaining complete control over data and costs.
