Open-source, self-hostable voice-agent platform with visual workflow builder, telephony, MCP-native control and BYO speech/model stack.
Product sources: Official Dograh site + GitHub repository
Dograh is an open-source voice-agent platform built around a visual workflow builder, telephony and a modular speech/model stack. Teams can use cascade pipelines with separate speech-to-text, language-model and text-to-speech providers or speech-to-speech models, while self-hosted and private-cloud deployments can keep the stack inside their own infrastructure. The current product also exposes Model Context Protocol control so compatible coding or agent tools can create and modify voice agents. 12
The open-source repository and official site describe configurable telephony, model providers, testing and workflow orchestration rather than a single bundled model. That makes Dograh closer to voice-agent infrastructure than to a finished conversational assistant. 12
Dograh's open-source self-hosted platform is free under the BSD 2-Clause license. Its managed cloud currently lists a 1¢ per minute platform fee, includes 10 concurrent calls and supports self-serve credits from $5; committed-volume and enterprise/on-premise plans use custom pricing. Model, speech and telephony provider costs can still sit underneath the platform fee depending on the chosen stack. 23
The combination of a free self-hosted edition and a low-friction managed option gives teams a practical way to evaluate the product before committing to private-cloud or enterprise deployment. 3
Dograh's strongest trust property is inspectability: the core platform is published under a permissive open-source license, and self-hosting or private-cloud deployment can keep audio, transcripts and model traffic inside infrastructure controlled by the adopter. Those deployment properties can support data-residency strategies, but they do not themselves mean Dograh is independently certified for every compliance framework named in its marketing. 12
A current independent Hlido review reports an exact-product hands-on assessment and gives Dograh 76/100, with four of five public-surface claims passing. It credits the depth of the feature surface and verifies open-source and traction signals, while noting that voice-quality and privacy claims were not fully tested on the public surface. That is useful independent context, but it remains a single editorial test rather than a broad external rating corpus. 4
Dograh is best suited to engineering-led voice-AI teams, contact-center builders and organizations with data-residency or infrastructure-control requirements. It is especially relevant when a team wants to choose its own STT, LLM, TTS and telephony components or run the system inside its own VPC or on-premise environment. 12
Teams that want a fully outsourced SaaS experience with minimal infrastructure ownership may prefer a more managed voice-agent vendor, although Dograh's managed cloud reduces that gap. 3
The main strengths are deployment control, open-source transparency and breadth across the voice-agent stack. BSD-2 licensing, self-hosting, private-cloud options, a visual builder, telephony, speech-to-speech support and MCP-native control give technical teams several paths from experimentation to production. Public pricing also makes the managed-cloud entry point unusually easy to compare. 123
The main weakness is maturity. Dograh is still a young platform, and many performance, scale and compliance-adjacent claims come from the vendor rather than a large independent evidence base. Self-hosting also transfers monitoring, upgrades, capacity planning and security operations to the adopter. The independent test is encouraging but not broad enough to establish long-term reliability across deployments. 14
Not separately stated in the source review.
Products matched by shared tasks, audience, workflow, product type and capabilities.