Bolna

What does Bolna do?

Bolna is an "end-to-end open-source voice agents platform" that lets you "quickly build voice firsts conversational assistants through a json", in its README's words. It "orchestrates voice conversations using combination of different ASR+LLM+TTS providers and models over websockets".

Telephony is Twilio and Plivo, with Exotel and Vonage marked coming soon. Transcription names Deepgram and Azure. Language models go through LiteLLM, so OpenAI, DeepSeek, Llama, Cohere, Mistral and others work, plus models served by vLLM. Text to speech lists AWS Polly, ElevenLabs, Deepgram, OpenAI, Azure, Cartesia and others. A text-only pipeline with no speech is also supported.

Key facts

• Licence: MIT, copyright Bolna, with no added conditions.
• The README's description: it "orchestrates voice conversations using combination of different ASR+LLM+TTS providers and models over websockets".
• Telephony: Twilio and Plivo, with Exotel and Vonage marked coming soon; a provider must support bi-directional streaming.
• Transcription: Deepgram and Azure named. Language models through LiteLLM (OpenAI, DeepSeek, Llama, Cohere, Mistral and others) or a vLLM server.
• Text to speech: AWS Polly, ElevenLabs, Deepgram, OpenAI, Azure, Cartesia, Smallest and others.
• A text-only pipeline with no speech is supported.
• Runs from the local setup folder with Docker Compose: a telephony web server, the Bolna server, ngrok and Redis; or imported in Python without telephony.
• The maintainers sell hosted APIs and a no-code playground built on this code, both closed source. Latest tagged release when read: 0.10.239.

What does it replace, and where does it fall short?

Bolna replaces Vapi and Retell AI for a phone agent defined in a file and run in four containers, where the models are hosted services you already have keys for.

Where it falls short, in the maintainers' own words: "We are actively looking for maintainers" and "We have fluctuated b/w maintaining this repository purely from a point of time crunch and not interest." The local example setup is marked for a move to another repository. Interruption handling and WebRTC are not described. The maintainers sell hosted APIs and a no-code playground built on this code, both closed source.

How does Bolna run?

Docker Compose from the local setup folder, starting a telephony web server (Twilio or Plivo), the Bolna server, ngrok and Redis. It can also be imported "directly in Python without the local telephony setup". A Twilio or Plivo account and an ngrok token are needed for calls. The code is MIT with no added conditions.

Who is Bolna for?

A small team that wants a phone agent running today from hosted providers, and is comfortable with a project whose maintenance its own README describes as uneven. For a larger build, Pipecat or LiveKit Agents.

What limits does the README state?

The maintainer capacity notes above; the pending move of the local setup; Exotel and Vonage not yet available.

Questions people ask

Is Bolna open source?

Yes. The repository is MIT with no added conditions, and the README says every integration lands in the open-source package first because it is the backbone of the maintainers' hosted APIs. The hosted APIs and the no-code playground are closed source.

Can Bolna make phone calls?

Yes, through Twilio or Plivo, which are the two carriers in its configuration table; Exotel and Vonage are marked coming soon. Calls need a carrier account and an ngrok token, and the local setup starts a telephony web server for the carrier you choose.

Is Bolna actively maintained?

Its README says, in its own words, "We are actively looking for maintainers" and that the team has "fluctuated b/w maintaining this repository purely from a point of time crunch and not interest". The repository had commits and a tagged release in September 2026 when read. Weigh that before building on it.

How does Bolna compare with Vapi?

Bolna is the fastest of the four frameworks on this site to a working phone agent: a JSON file and one Docker Compose command. Vapi is a hosted platform with a dashboard. For a larger build, Pipecat or LiveKit Agents; the alternatives to Vapi page compares all four.

Which language models does Bolna support?

Any model LiteLLM can reach, which the README lists as OpenAI, DeepSeek, Llama, Cohere, Mistral and others, plus models served by a vLLM server through a base URL. The choice is made in the agent's JSON definition.

Sources

• Bolna README and MIT LICENSE: github.com/bolna-ai/bolna, read 2026-09-16.