Dograh
7/10Dograh is an open-source, self-hostable platform for building inbound and outbound AI phone agents with a visual workflow builder and bring-your-own models.
Data last updated November 2026 · Our methodology
Overview
Dograh is a voice AI platform under the BSD-2 license. It is positioned as an open alternative to Vapi and Retell. Teams design phone agents in a visual, node-based workflow builder that includes agent nodes, global instructions, tool calls, transitions, QA nodes, knowledge bases and end-call outcomes. Agents can handle inbound and outbound calls, run either an LLM/STT/TTS pipeline or a speech-to-speech model, and hand calls to a human with a warm transfer. Telephony connectors cover Twilio, Vonage, Telnyx, Plivo, Cloudonix, Vobiz, Exotel and Asterisk ARI. You can bring keys for providers such as OpenAI, Deepgram, ElevenLabs and Cartesia, or run local models through Ollama, vLLM or Speaches. Developers get Python and Node SDKs, webhooks, and an MCP server that lets Claude Code, Cursor or Codex build and edit agents. Pricing has four options. The self-hosted open-source edition is free with no limits. Pay as You Go cloud charges a 1 cent per minute platform fee plus provider usage, with credits starting at $5. Custom Volume and Enterprise plans are quoted on request, and Enterprise adds on-premise deployment, SSO/SAML and audit logs. The product suits developers, startups and agencies that want control over cost and data. Its main limitation is maturity. The repository dates from September 2025, independent hands-on reviews are scarce, and support on the free tier runs mostly through GitHub and Slack. Self-hosting also needs Docker and some infrastructure skills.
Our Rating
Who Dograh is best for
Best software for DevelopersGood fit
It offers open-source code, SDKs, an MCP server and swappable model providers.
It is free to self-host, and managed cloud costs 1 cent per minute with no commitment.
Agencies can fork the stack and build phone agents for several clients without vendor lock-in.
Less suited
Inbound call handling and warm transfer to humans are supported, but setup takes technical work.
On-premise deployment, SSO and audit logs are available on the Enterprise plan, but the vendor is young.
Pros & Cons
Pros
- The full platform is free to self-host under the permissive BSD-2 license, with no usage limits.
- The managed cloud platform fee of 1 cent per minute is low, and provider costs are passed through via BYOK.
- It connects to many telephony providers, including Twilio, Vonage, Telnyx, Plivo and Asterisk ARI.
- It can run fully local using Ollama, vLLM and Speaches, which suits data-sensitive deployments.
- The MCP server and the Python and Node SDKs let developers build agents from coding assistants.
Cons
- It is a young project, with the repository created in September 2025, and there are few independent reviews.
- Self-hosting needs Docker, at least 8 GB of RAM and 4 vCPUs, and open telephony ports.
- LLM, STT, TTS and telephony costs are billed separately, which makes total cost harder to predict.
- Support on the free tier is mainly through GitHub Issues and the Slack community.
Key Features
Dograh Pricing
Open Source
- BSD-2 license
- No usage limits
- Bring your own API keys
- Self-hosted full platform
Custom Volume
- Committed monthly volume discounts
- Priority support via email, Slack or WhatsApp
Enterprise
- SSO/SAML
- Audit logs
- Custom SLAs
- Data residency
- Dedicated account manager
- On-premise or private cloud
Pay as You Go
- Credits from $5
- 10 concurrent calls
- 1 cent per minute platform fee
- BYOK or Dograh Models with usage pricing
Frequently Asked Questions
Yes. The open-source edition can be self-hosted for free with no limits. You pay only your own provider costs.
Pay as You Go charges a 1 cent per minute platform fee plus model and telephony usage. Credits start at $5, and the plan includes 10 concurrent calls.
Twilio, Vonage, Telnyx, Plivo, Cloudonix, Vobiz, Exotel and Asterisk ARI.
Yes. It deploys with Docker. The Enterprise plan adds on-premise or private-cloud deployment with data residency, SSO/SAML and audit logs.
Anonymous usage data is collected by default. You can turn it off by setting ENABLE_TELEMETRY=false.
Similar Tools
See all alternatives
AI meeting assistant that records, transcribes and summarizes Zoom, Google Meet and Teams calls, then pushes notes and insights to CRMs and work tools.
Langdock is a Berlin-based, EU-hosted AI workspace where company teams chat with many AI models, build custom agents, and automate workflows under central governance.
LangChain offers open-source frameworks (LangChain, LangGraph, Deep Agents) plus the paid LangSmith platform. Developers use them to build, evaluate, deploy and monitor LLM agents.
Ranked near Dograh in AI Software
See Dograh in the listDograh is #69 in AI Software by our score.
