Apache-2.0 open-source platform for creating voice agents across web, phone, WhatsApp and Telegram with dashboard, SDK, CLI and self-hosting.
Product sources: Official GitHub repository
OpenVox is an Apache-2.0 open-source platform for building voice agents through a local dashboard, CLI and REST/WebSocket API. Its current repository combines a voice pipeline, 33 stated templates, scheduling, observability and evaluation tools with web, phone, WhatsApp and Telegram channels. Agents can run a conventional speech-to-text → model → text-to-speech pipeline or a speech-to-speech path, with provider choice across several model and speech vendors. 1
The platform is designed to be usable at two levels: non-technical operators can copy and configure templates in the dashboard, while developers can add Python skills, use the HTTP API or work with the repository's SDK code. The repository explicitly notes that standalone TypeScript and Python SDK packages are not yet published, so the REST/WebSocket interface is the stable integration path today. 1
OpenVox itself is free to self-host under the Apache 2.0 license. No separate paid OpenVox platform or public subscription pricing was verified. Users bring their own model, speech, telephony and messaging provider credentials, so those underlying services can still create usage costs. 12
The core Python package is distributed through PyPI and requires Python 3.11 or newer; Docker and local installation paths are also documented. 12
The product's strongest trust property is source availability. The repository exposes the implementation, Apache license, deployment configuration and data-handling controls. It states that OpenVox itself has no telemetry, stores provider keys encrypted locally, and supports configurable retention, transcript-only operation and PII masking. These are inspectable first-party implementation claims rather than independent privacy certification. 1
PyPI independently confirms the openvox-core package identity, Apache licensing, Python requirement and beta development classifier. It supports the existence and distribution of the implementation but does not validate latency, template quality, reliability or privacy behavior. 2
OpenVox is most relevant to developers, automation teams and self-hosters who want one local platform spanning browser voice, phone calls and messaging channels while retaining provider and infrastructure control. Its dashboard and templates may also appeal to less technical operators working inside a team that can own the deployment. 1
It is less suitable today for organizations that require a mature managed service, established enterprise support, large-scale adoption evidence or fully packaged SDKs across every promised development surface. 12
The main strength is scope. A single open-source project covers dashboard configuration, agent templates, two voice architectures, provider switching, telephony and messaging channels, scheduling, observability, evaluation, storage choices and an extensibility layer. The REST/WebSocket API also gives developers a usable integration surface before the standalone SDK packages ship. 1
The main weakness is maturity. The public repository currently shows very limited adoption, PyPI labels the package beta, the standalone TypeScript/Python SDK packages are not published, and some advertised channel adapters are explicitly scaffolded or awaiting real-account testing. Performance figures such as latency and rapid setup remain first-party claims, so production teams should validate them against their own providers and traffic before relying on them. 12
Not separately stated in the source review.
Products matched by shared tasks, audience, workflow, product type and capabilities.