dograh-hq/dograh

📚 Summary

Dograh AI is an open‑source, self‑hostable platform for building voice AI agents. It lets you create, test, and run phone‑based conversational bots through a visual workflow editor. The stack is modular – you can plug in any LLM, speech‑to‑text (STT), text‑to‑speech (TTS) or telephony provider, or use the bundled stack that ships with the product. A one‑command Docker setup gets a full instance running in minutes, and a cloud‑hosted SaaS version is also offered.

🛠️ Core Features

Feature	What it does
Visual workflow builder	Drag‑and‑drop nodes (start, agent, tools, QA, knowledge bases, webhooks, end) to design call flows without coding.
Test panel	Test Audio lets you speak to the bot in the browser; Test Chat enables fast text‑based iteration and replay of conversation turns.
Telephony integrations	Built‑in connectors for Twilio, Vonage, Telnyx, Plivo, Vobiz, Cloudonix, Asterisk ARI, plus human hand‑off via call transfer.
Pluggable AI stack	Bring your own LLM, STT, TTS, or use Dograh’s default stack; keys are configurable via environment variables.
MCP (Machine Coding Protocol) support	Connect Claude Code, Codex, Cursor or any MCP client to programmatically edit workflows, fetch schemas, and generate agents from natural‑language prompts.
SDKs	Python (dograh-sdk) and Node (@dograh/sdk) libraries for creating agents and making outbound calls programmatically.
One‑command Docker deployment	curl … && ./start_docker.sh pulls a docker‑compose.yaml and launches the whole stack (API, UI, MinIO storage, optional Cloudflare tunnel).
Self‑hosted data	All data lives on your infrastructure (MinIO or S3‑compatible storage), giving full data residency control.

🚀 Getting Started (quick‑start)

# Grab the Docker compose file and starter script
curl -o docker-compose.yaml https://raw.githubusercontent.com/dograh-hq/dograh/main/docker-compose.yaml \
     -o start_docker.sh https://raw.githubusercontent.com/dograh-hq/dograh/main/scripts/start_docker.sh
chmod +x start_docker.sh
# Launch the platform (creates a public Cloudflare tunnel for inbound webhooks)
./start_docker.sh

• After a few minutes the UI is reachable at http://localhost:3010.
• Create an account, then use the visual builder to define a voice bot.
• Test it instantly with the built‑in audio or chat tester.

📦 Architecture Overview (high‑level)

• Frontend – React UI served on port 3010, provides the workflow editor and testing tools.
• Backend – Python service handling API requests, orchestrating LLM calls, STT/TTS, and telephony webhook routing.
• Storage – MinIO (self‑hosted S3‑compatible) or external S3 for persisting recordings, prompts, and bot definitions.
• Telephony gateway – Connectors translate incoming/outgoing call events from providers into a unified internal format.
• MCP server – Exposes an endpoint that coding assistants (Claude Code, Codex, etc.) can call to read/write workflow JSON.

🤝 Community & Support

• Slack community – active channel for troubleshooting, feature discussion, and onboarding.
• GitHub Discussions & Issues – place for use‑case sharing, bug reports, and PRs.
• Documentation – full docs at https://docs.dograh.com covering deployment, MCP integration, and SDK usage.

📜 License

BSD 2‑Clause – permissive, allowing commercial use, modification, and redistribution.

---

All information above is taken directly from the repository’s README; no external assumptions have been added.