Back to catalog
Fonoster
MITFonoster is an open-source alternative to Twilio for building programmable voice apps and IVRs with Node.js and self-hosted SIP infrastructure.
Open-source alternative to — compare it against:
Overview
Pedro Sanders originally developed Fonoster as a proprietary telecommunications platform in 2016, releasing it as open source in late 2020 after seeing how closed carrier monopolies constrained modern developers. The project strips away proprietary voice layers, turning raw SIP signaling and media processing into clean Node.js APIs. Today, it functions as a self-hosted telecommunications utility that isolates call handling logic from commercial carrier contracts.
Why Fonoster instead of Twilio
Twilio bundles application code, SIP proxies, and a global PSTN carrier network into a unified managed service. That convenience comes with recurring per-minute voice fees, platform markups on phone numbers, and complete dependency on their proprietary cloud. You write webhooks against their API and avoid maintaining telephony infrastructure, but your operational spending scales directly with call volume, and customer audio travels across external infrastructure.
Fonoster separates the programmable software layer from carrier connectivity. Teams run the voice application engine—complete with verbs like Answer, Play, Say, Gather, and Dial—on private servers while connecting to wholesale SIP trunking providers of their own choice. Fonoster does not vend phone numbers or operate an internal carrier network, so ops teams take on SIP routing, NAT traversal, and container orchestration. In exchange, the architecture removes vendor markups, keeps call recordings within local storage boundaries, and allows switching voice carriers without touching application logic.
What it does well
- Bring-your-own-carrier SIP trunking: route calls through any wholesale SIP provider, such as Telnyx or VoIP.ms, without paying software surcharges on top of standard carrier minutes.
- Programmable voice applications: control inbound and outbound call flows in TypeScript or JavaScript using straightforward verbs instead of writing raw Asterisk dial plans.
- Integrated speech synthesis and recognition: connect directly to Google Speech APIs to build interactive voice response menus and automated voice agents without separate middleware.
- Multi-tenant architecture: organize domains, numbers, and access control lists so agencies can host distinct client telephone systems on one shared cluster.
