Sarvam AI - Bruno API Collection

A ready-to-use Bruno collection for the Sarvam AI APIs — Text-to-Speech (TTS) and Speech-to-Text (STT) across 23 Indian languages.

Prerequisites

• Bruno installed on your machine
• A Sarvam AI API key

Getting Your API Key

1. Go to the Sarvam AI Dashboard
2. Sign up or log in to your account
3. Navigate to the API Keys section and create a new key
4. Copy your API key — you'll need it when configuring the Bruno environment

For more details, see the Sarvam AI Quickstart Guide.

Importing the Collection in Bruno

1. Open Bruno
2. Click the + button in the top-left corner and select Import Collection
3. Select Clone from Git Repository
4. Paste the repository URL:

   https://github.com/bruno-collections/bruno-sarvamai.git
   

5. Choose a local destination folder
6. Click Clone

The collection will be cloned and loaded with all TTS and STT requests organized by language.

For more details on importing, see the Bruno Import Collections Guide.

Configuring the API Key

1. After importing, select the Sarvam-AI collection in Bruno
2. Go to Environments and select the local environment
3. Set the value of Sarvam_API_Key to your Sarvam AI API key
4. Make sure the local environment is active

Folder Structure

collections/Sarvam-AI/
├── opencollection.yml            # Collection metadata
├── .gitignore
├── environments/
│   └── local.yml                 # Environment with Sarvam_API_Key secret
├── text-to-speech/
│   ├── folder.yml                # Folder config
│   ├── text-to-speech.yml        # Original TTS request (Hindi)
│   ├── tts-bengali.yml           # TTS → Bengali (bn-IN)
│   ├── tts-english.yml           # TTS → English (en-IN)
│   ├── tts-gujarati.yml          # TTS → Gujarati (gu-IN)
│   ├── tts-hindi.yml             # TTS → Hindi (hi-IN)
│   ├── tts-kannada.yml           # TTS → Kannada (kn-IN)
│   ├── tts-malayalam.yml         # TTS → Malayalam (ml-IN)
│   ├── tts-marathi.yml           # TTS → Marathi (mr-IN)
│   ├── tts-odia.yml              # TTS → Odia (od-IN)
│   ├── tts-punjabi.yml           # TTS → Punjabi (pa-IN)
│   ├── tts-tamil.yml             # TTS → Tamil (ta-IN)
│   ├── tts-telugu.yml            # TTS → Telugu (te-IN)
│   ├── tts-assamese.yml          # TTS → Assamese (as-IN) [Beta]
│   ├── tts-bodo.yml              # TTS → Bodo (brx-IN) [Beta]
│   ├── tts-dogri.yml             # TTS → Dogri (doi-IN) [Beta]
│   ├── tts-konkani.yml           # TTS → Konkani (kok-IN) [Beta]
│   ├── tts-kashmiri.yml          # TTS → Kashmiri (ks-IN) [Beta]
│   ├── tts-maithili.yml          # TTS → Maithili (mai-IN) [Beta]
│   ├── tts-manipuri.yml          # TTS → Manipuri (mni-IN) [Beta]
│   ├── tts-nepali.yml            # TTS → Nepali (ne-IN) [Beta]
│   ├── tts-sanskrit.yml          # TTS → Sanskrit (sa-IN) [Beta]
│   ├── tts-santali.yml           # TTS → Santali (sat-IN) [Beta]
│   ├── tts-sindhi.yml            # TTS → Sindhi (sd-IN) [Beta]
│   └── tts-urdu.yml              # TTS → Urdu (ur-IN) [Beta]
├── speech-to-text/
│   ├── folder.yml                # Folder config
│   ├── speech-to-text.yml        # Original STT request
│   ├── echo-sarvam.yml           # Translate API request
│   ├── stt-assamese.yml          # STT ← Assamese (as-IN)
│   ├── stt-bengali.yml           # STT ← Bengali (bn-IN)
│   ├── stt-bodo.yml              # STT ← Bodo (brx-IN)
│   ├── stt-dogri.yml             # STT ← Dogri (doi-IN)
│   ├── stt-english.yml           # STT ← English (en-IN)
│   ├── stt-gujarati.yml          # STT ← Gujarati (gu-IN)
│   ├── stt-hindi.yml             # STT ← Hindi (hi-IN)
│   ├── stt-kannada.yml           # STT ← Kannada (kn-IN)
│   ├── stt-kashmiri.yml          # STT ← Kashmiri (ks-IN)
│   ├── stt-konkani.yml           # STT ← Konkani (kok-IN)
│   ├── stt-maithili.yml          # STT ← Maithili (mai-IN)
│   ├── stt-malayalam.yml         # STT ← Malayalam (ml-IN)
│   ├── stt-manipuri.yml          # STT ← Manipuri (mni-IN)
│   ├── stt-marathi.yml           # STT ← Marathi (mr-IN)
│   ├── stt-nepali.yml            # STT ← Nepali (ne-IN)
│   ├── stt-odia.yml              # STT ← Odia (od-IN)
│   ├── stt-punjabi.yml           # STT ← Punjabi (pa-IN)
│   ├── stt-sanskrit.yml          # STT ← Sanskrit (sa-IN)
│   ├── stt-santali.yml           # STT ← Santali (sat-IN)
│   ├── stt-sindhi.yml            # STT ← Sindhi (sd-IN)
│   ├── stt-tamil.yml             # STT ← Tamil (ta-IN)
│   ├── stt-telugu.yml            # STT ← Telugu (te-IN)
│   └── stt-urdu.yml              # STT ← Urdu (ur-IN)
└── output/
    ├── folder.yml                # Folder config
    └── *.wav                     # Generated audio output files

Supported Languages

Language Code	Language	TTS	STT	Status
bn-IN	Bengali	✅	✅	Stable
en-IN	English	✅	✅	Stable
gu-IN	Gujarati	✅	✅	Stable
hi-IN	Hindi	✅	✅	Stable
kn-IN	Kannada	✅	✅	Stable
ml-IN	Malayalam	✅	✅	Stable
mr-IN	Marathi	✅	✅	Stable
od-IN	Odia	✅	✅	Stable
pa-IN	Punjabi	✅	✅	Stable
ta-IN	Tamil	✅	✅	Stable
te-IN	Telugu	✅	✅	Stable
as-IN	Assamese	✅	✅	Beta
brx-IN	Bodo	✅	✅	Beta
doi-IN	Dogri	✅	✅	Beta
kok-IN	Konkani	✅	✅	Beta
ks-IN	Kashmiri	✅	✅	Beta
mai-IN	Maithili	✅	✅	Beta
mni-IN	Manipuri	✅	✅	Beta
ne-IN	Nepali	✅	✅	Beta
sa-IN	Sanskrit	✅	✅	Beta
sat-IN	Santali	✅	✅	Beta
sd-IN	Sindhi	✅	✅	Beta
ur-IN	Urdu	✅	✅	Beta

Note: Languages marked as Beta are in beta access and may have limited support or accuracy compared to stable languages. Refer to the Sarvam AI documentation for the latest availability.

How It Works

Text-to-Speech (TTS)

Each TTS request sends a sample text in the target language to the Sarvam AI /text-to-speech endpoint using the bulbul:v3 model. The after-response script automatically decodes the base64 audio and saves it as a .wav file in the output/ folder (e.g., output/tamil-output.wav).

Speech-to-Text (STT)

Each STT request uploads an audio file to the Sarvam AI /speech-to-text endpoint using the saaras:v3 model in transcribe mode. The file field expects an audio file named <language>.wav (e.g., bengali.wav). Add your audio files before running the requests.

API Reference

• Sarvam AI API Docs
• Text-to-Speech API
• Speech-to-Text API
• Supported Models

License

This collection is provided as-is for use with Sarvam AI APIs. You must have a valid Sarvam AI subscription to use these endpoints.