Choosing a speech-to-text provider

The transcriber turns the caller’s audio into text. Language support is the primary selection factor — most transcribers are optimized for specific language families.

Provider	Best for	Notes
Deepgram Nova-3	English	Fastest; best accuracy for English
Deepgram Nova-2	Many languages	Broader language coverage
ElevenLabs	English	High accuracy; slightly higher latency
Azure Speech	Enterprise, many languages	Strong multilingual; good for EU
AssemblyAI	English	High accuracy; async features
Sarvam	Indian languages	Best for Hindi, Tamil, Bengali, Telugu, Marathi, etc.
Gladia	European languages	Good multilingual coverage
Soniox	Code-switching	One model auto-detects and switches language mid-sentence
Smallest AI	Ultra-low latency	Pulse model; sub-100ms first-token latency
Gemini 3.5 Transcribe	Multilingual	High accuracy and low latency across many languages
Deepgram Flux	Lowest latency	Turn detection inside the model; enables speculative LLM responses and instant barge-in
OpenAI Realtime	Clean or accented audio	Built-in VAD; `delay` trades latency against accuracy
Google	Broad language coverage	Google Cloud Speech-to-Text

Key settings: endpointing (silence detection), language (always set explicitly — auto-detection adds latency), encoding / sampling_rate (must match your telephony provider). For the exact provider and model values each takes, see Create Agent.

Next steps

Speech-to-text providers

Setup guides for every transcriber Bolna supports

Supported languages

Languages and codes you can configure

Choose an LLM

The next provider in the pipeline

Back to the decision guide

Compare all provider categories side by side

Telephony

AI models

No-code automation

Choose providers

Choosing a speech-to-text provider

Next steps

Speech-to-text providers

Supported languages

Choose an LLM

Back to the decision guide

​Next steps

Speech-to-text providers

Supported languages

Choose an LLM

Back to the decision guide

Next steps