Deepgram
Plug Deepgram speech-to-text into Bolna for fast, accurate phone call transcription. Supports low latency streaming, high accuracy, and custom vocabulary.
How Deepgram fits in the stack
Best for
Turning noisy phone audio into text the agent can reason over.
Use this layer when
Recognition accuracy, language coverage, and latency matter most.
Connects to
Telephony audio upstream and your LLM downstream.
Speech-to-text is the listening layer
This provider converts raw audio into text in real time. It shapes how accurately the agent hears the caller and how natural the back-and-forth feels.
Workflow
Telephony Phone Network → STT Listener → LLM Reasoning → TTS Voice → Tools Actions
This page focuses on where Deepgram fits in a production voice stack. For full setup steps, credentials, and API details, use the documentation link above.
Overview
Deepgram is the leading speech-to-text (STT) provider purpose-built for real-time voice applications. With industry-leading accuracy and ultra-low latency, Deepgram powers the listening capability of your Bolna voice agents, converting spoken words into text that your LLM can understand and respond to.
Deepgram's AI models are trained on diverse datasets including phone conversations, making it the ideal choice for voice agent applications where every word counts.
Features & Use Cases
Highly Accurate (Nova-2 Model)
Deepgram gets it right over 95% of the time, easily understanding different accents, background noise, and varying phone call qualities.
Lightning-Fast Speed
It turns speech into text in under 300 milliseconds. This real-time speed prevents awkward pauses and keeps conversations smooth.
Custom Words & Jargon
You can teach the AI your specific industry terms, product names, or technical jargon so it always recognizes them correctly.
Speaks Many Languages
Understands over 30 languages - including English, Spanish, French, German, and Hindi - and accurately handles native accents.
Clean, Readable Text
Automatically adds punctuation, capitalizes the right words, and formats numbers so the final text is clean and ready to use.
Optimized for Phone Calls
Specially trained to handle typical phone call audio perfectly, making it ideal for call centers and customer support.
Knows Who is Speaking
Can automatically figure out and separate who is talking at any given moment, which is perfect for distinguishing the agent from the customer.
Other providers in this layer
Keep exploring the voice stack
See where Deepgram fits in your production workflow
Use the demo to walk through provider selection, stack tradeoffs, and the exact workflow you want Bolna to automate.