Languages Tab showing per-language Voice (ElevenLabs) and Transcription (Deepgram) settings
Languages
The strip at the top of the tab, labeled Settings for, selects which language the voice and transcription settings below apply to. The language marked(Default) is the one your agent opens every conversation in.
Settings for strip with Hindi as the default language and English as the additional language, and the switching-condition dropdown below
Languages are added, removed, and promoted to default in the Agent Tab. On this tab the strip only switches between the languages you already configured there.
When the agent switches language
For multilingual agents, a summary under the strip recaps the setup — for example “Starts in English, can move to Hindi, Dutch when …” — and ends in a dropdown that controls when a switch is allowed: requested or auto detected (the default) or the caller requested for it. See When the Agent May Switch for what each option means, and How Language Switching Works for exactly what counts as “the caller speaking another language” — why a stray English word or a read-out order number won’t trip a switch.Supported languages
See the Supported languages page for the full list of languages Bolna can be configured with.Voice
Controls how your agent sounds when speaking to the caller. For multilingual agents, each language can have its own voice provider, model, and voice. Select a language in the strip above to configure its voice settings independently.
Voice configuration with ElevenLabs as provider, Eleven Multilingual v2 as model, and the Tripti voice selected, with the Advanced settings panel collapsed
Provider, model, and voice
1
Select a Provider
Choose from AzureTTS, Cartesia, ElevenLabs, or Sarvam.
2
Pick a Model
Select the model that fits your latency and quality needs (e.g., ElevenLabs
eleven_turbo_v2_5 for low latency).3
Choose a Voice
Click the Voice dropdown to browse all voices for the selected provider and model.
Browsing voices
Click the Voice dropdown to see a searchable list of all available voices. Filter by gender using the All, Male, Female, and Neutral tabs. Each voice shows a play button so you can preview it before selecting.
Voice selector dropdown showing a searchable list of voices with gender filter tabs and play preview buttons
Preview welcome message
Click the play button next to the Voice dropdown to hear the selected voice speak your agent’s welcome prompt (configured in the Agent Tab). This lets you test how the voice sounds before going live.Voice tuning parameters
Fine-tune your agent’s voice using the sliders under Advanced settings, below the voice selector. Available parameters may vary by provider.Adding and cloning voices
Open the Voice dropdown and click Add new voice at the bottom to add a custom voice by ID or clone one from an audio sample. This opens the Add Voice dialog.Custom voice uploads are only available for ElevenLabs and Cartesia.
Add a voice by ID
Use this when you already have a voice ID from your provider’s voice library.
Add Voice dialog with the Add by ID tab selected, ElevenLabs as provider, and a Voice ID input field
1
Select the Add by ID tab
Make sure the Add by ID tab is selected in the dialog.
2
Choose a Provider
Select ElevenLabs or Cartesia.
3
Enter the Voice ID
Paste the voice ID from your provider. For ElevenLabs, find IDs in the ElevenLabs voice library.
4
Click Add voice
The voice will appear in the Voice dropdown for all your agents.
Clone a voice
Create a new voice by uploading an audio recording. Useful for maintaining a consistent brand voice or using a specific person’s voice (with their permission).
Clone Voice tab with fields for Provider, Voice name, Description, Sample language, and a file upload area
1
Select the Clone Voice tab
Switch to the Clone Voice tab in the dialog.
2
Choose a Provider
Select ElevenLabs or Cartesia.
3
Enter Voice Details
Add a Voice name (e.g., “Sales Assistant Voice”) and Description (e.g., “Warm male Indian accent”).
4
Select Sample Language
Choose the language of your audio sample.
5
Upload an Audio Sample
Drag and drop your audio file or click click to browse. Audio files only, maximum 10 MB.
6
Click Clone voice
The platform processes your sample and adds the new voice to the Voice dropdown.
Supported languages for voice cloning
Both ElevenLabs and Cartesia support cloning in the same languages Bolna supports, plus an additional Indian Multilingual sample-language option.Transcription
Controls how your agent converts the caller’s spoken words into text before the LLM processes them. For multilingual agents, each language can have its own transcription provider and model. Select a language in the strip above to configure its transcription settings independently.Different languages may perform better with different providers — see Per-Language Configuration for a worked example.
Transcription configuration with Deepgram nova-3 selected, plus Keywords and Context fields
Provider and model
Choose a transcription provider from the Provider dropdown, then pick the specific model from the Model dropdown.Keywords
Boost recognition accuracy for specific words the transcriber might miss, such as brand names, product names, or technical terms. Enter one term per entry, comma-separated (e.g.,Bruce, Acme, Meesho).
Not every transcriber reads this field,When your
selected transcriber or model can’t use it, the box is shown disabled. See
provider support for which providers
read it and their limits.
Advanced settings
Some transcribers have settings of their own, under Advanced settings below the Keywords and Context fields. The section is collapsed by default and only appears for a provider that has such settings.
AssemblyAI transcription with Advanced settings expanded, showing Mode, VAD threshold and Longest pause in a turn at AssemblyAI's defaults
Each AssemblyAI setting opens on the value AssemblyAI uses when you leave it alone, so you can see what you are changing from. Raise Longest pause in a turn if callers get cut off while reading out numbers or addresses, and raise VAD threshold if line noise interrupts the agent. See AssemblyAI session settings for what each does and the settings available over the API.
Hand-in
Shown only when the agent has more than one language. These two fields cover how the agent moves into the language selected in the strip above.Both are optional — leave them blank and the agent switches without a spoken hand-in line. These fields used to live under Advanced Settings in the Agent Tab, where they were called Agent Name and Handoff Message. They are still per language — the values set for Hindi do not affect English.
Next steps
Engine tab
Configure interruption handling, endpointing, and latency
Multilingual support
Set up agents that speak multiple languages in a single call
How language switching works
The detection engine behind mid-call switching
Clone voices
Create a custom voice from an audio sample
Deepgram provider
Explore Deepgram transcription models and keyword boosting