What is Dograh?
Dograh is an open-source, self-hostable voice agent platform designed as a powerful alternative to Vapi and Retell. It empowers developers to build, deploy, and manage voice agents with complete control over their data and infrastructure. The platform features a visual workflow builder that allows you to configure every component of your voice agent pipeline, including inbound channels, speech-to-text (STT), large language models (LLM), text-to-speech (TTS), and telephony. You can also bypass the cascade entirely and switch to a speech-to-speech pipeline for ultra-low latency and natural conversations.
One of Dograh's standout features is its data sovereignty. You can deploy it on-premises or within your own cloud VPC, and even serve AI models within your perimeter or in an air-gapped environment. This ensures that no calls, recordings, transcripts, or model inferences ever leave your boundary, making it ideal for regulated industries that require strict compliance and data residency. The platform is fully auditable, open source, and enterprise-ready, so your existing certifications and controls remain valid without needing to audit a third-party vendor.
Dograh also includes a Model Context Protocol (MCP) server, enabling agent runtimes like Claude Code, OpenClaw, Cursor, and Codex to spin up, modify, and deploy voice agents directly from your IDE. This integration streamlines development workflows and accelerates iteration.
For voice quality, Dograh combines real human voice clips with TTS in the same cloned voice. The LLM selects pre-recorded lines when appropriate and falls back to TTS only when needed, resulting in latency gains of up to 3x and cost reductions of up to 3x while sounding truly human. The audio-in, audio-out approach eliminates the transcription cascade, providing fluid turn-taking and interruption handling.
Use cases include customer support, virtual assistants, and any application requiring natural voice interactions with strict data privacy. Dograh supports models like Gemini 3.1 Flash Live, GPT Realtime 2, Voxtral, Kokoro, and Whisper, with STT and TTS production-ready and LLM swapping in beta.
Use Cases
Key Features
open source self hostable visual workflow builder speech to speech pipeline data sovereignty on premises deployment air gapped environment auditable enterprise readyWho is it for?
developers, regulated industries, compliance teams, enterprise IT, voice agent builders, privacy, conscious organizations
