Voice AI agents platform for building conversational apps
Top 81.1% on sourcepulse
Bolna provides an end-to-end, open-source framework for building LLM-powered voice AI agents. It orchestrates conversations by integrating various Automatic Speech Recognition (ASR), Large Language Model (LLM), and Text-to-Speech (TTS) providers via WebSockets, enabling developers to quickly create voice-first conversational assistants.
How It Works
Bolna acts as an orchestration layer, managing the flow of data between different AI services. It leverages a modular design to support a wide array of providers for ASR (e.g., Deepgram), LLMs (e.g., OpenAI, Mistral via LiteLLM), and TTS (e.g., AWS Polly, ElevenLabs). The platform handles real-time audio streaming and conversation state management, allowing for flexible and scalable voice agent development.
Quick Start & Requirements
docker-compose build --no-cache bolna-app <twilio-app | plivo-app>
and run with docker-compose up bolna-app <twilio-app | plivo-app>
.DEEPGRAM_AUTH_TOKEN
, OPENAI_API_KEY
, TWILIO_ACCOUNT_SID
).Highlighted Details
Maintenance & Community
Licensing & Compatibility
Limitations & Caveats
The README notes that hosted APIs and the no-code UI playground are currently closed source. The project is actively looking for maintainers, which may indicate a potential risk to ongoing development velocity.
6 days ago
1 day