Back to list

Dograh

An open-source, self-hostable voice AI platform — build production voice agents with a visual workflow builder, your own models, your own keys.

袁知夏
Author
@袁知夏
Project ShowcaseDeveloper ToolsOnlineclaude-code
0
0

Project intro

Dograh is an open-source, self-hostable platform for building production voice agents — an alternative to Vapi and Retell. You draw the agent's workflow in a visual builder, test it by voice or text in the browser, and put it on the phone. Run it on your own server with one Docker command, or use the hosted cloud.

Workflows are built from agent nodes, tools, transitions, and end-call outcomes, with a QA node, knowledge bases, and webhooks for production use. Bring your own language model, speech-to-text, text-to-speech, and phone provider — Twilio, Vonage, Telnyx, and more — or use the built-in stack with no keys at all. Coding agents can build and edit workflows through an MCP server, and there are Python and Node SDKs for creating agents and placing calls in code.

More than 6,000 developers starred it, with nearly 1,600 forks, and version 1.51.0 shipped on 4 November 2026. It was Product Hunt's #1 Product of the Day and #1 Product of the Week. It is BSD-2-Clause licensed.

Two honest limits: self-hosting still needs a server of your own plus a phone provider before real calls work, and the hosted cloud is usage-based rather than flat-rate. The quickstart collects anonymous usage data, which you can switch off.

Actions

Visit project URL

Tech stack

python

Project info

Project category
Developer Tools
Project status
Online
Collaboration intent
None

Comments (0)

ME
0/1000
Loading comments...