Dograh

// Overview

Dograh is an open-source platform for building production voice agents, positioned as a self-hostable alternative to the hosted voice-agent services. You design a conversation as a visual workflow, test it in the browser, and deploy it against inbound or outbound telephony — with every line of the stack available to modify.

Its defining choice is where the model boundary sits. Dograh is bring-your-own-key across the whole pipeline: run a single speech-to-speech model, or compose separate LLM, STT and TTS providers, or point it at a local model. Because it self-hosts with one Docker command, the audio and the transcripts stay on your infrastructure rather than passing through a vendor's cloud — the practical reason teams in regulated settings reach for it.

The platform is MCP-native, which means AI coding assistants can help design and edit the workflows themselves: the project ships an official MCP integration for Claude Code and Codex so an agent can build the voice agent. It is built on Pipecat for the real-time media pipeline, supports Asterisk ARI for telephony, and is released under the BSD 2-Clause licence by a team of YC alumni and exit founders.

// What it does

• Visual workflow builder for designing and testing voice agents without writing the conversation logic by hand
• Bring your own LLM, STT and TTS — or a single speech-to-speech model — including local models, with no provider lock-in
• MCP-native: an official MCP server lets Claude Code or Codex design and edit workflows for you
• Telephony support for inbound and outbound calls, including Asterisk ARI
• Self-hosts with one Docker Compose command, on-prem or on a remote server, so audio and transcripts stay on your infrastructure
• BSD 2-Clause licensed with the full source open, alongside an optional hosted cloud

// Getting started

Dograh self-hosts via Docker Compose. One script fetches the compose file and starts the stack; a hosted cloud is available at app.dograh.com if you want to try it without installing anything.

curl -o docker-compose.yaml https://raw.githubusercontent.com/dograh-hq/dograh/main/docker-compose.yaml \
  && curl -o start_docker.sh https://raw.githubusercontent.com/dograh-hq/dograh/main/scripts/start_docker.sh \
  && chmod +x start_docker.sh && ./start_docker.sh

Commands and code are distilled from the project's own documentation — always check the official repo for the latest.

// When to use it

• Run inbound or outbound call agents on your own infrastructure when call audio cannot leave your network
• Replace a per-minute hosted voice platform with a self-hosted stack you can modify at source level
• Compose a voice pipeline from the specific LLM, STT and TTS vendors you have already chosen, rather than the ones a platform supports
• Prototype a voice agent quickly in the visual builder, then hand the workflow to a coding assistant over MCP for refinement

// How Dograh compares

Dograh alongside other open-source audio, music & voice tools AI/TLDR tracks, ranked by GitHub stars.