![](data:image/svg+xml,<svg display="block" role="presentation" viewBox="0 0 107 20" xmlns="http://www.w3.org/2000/svg"><g d="M 92.359 20 L 90.445 20 L 97.72 0 L 99.673 0 L 106.948 20 L 105.034 20 L 98.764 2.373 L 98.628 2.373 L 92.358 20 Z M 87.012 0 L 87.012 20 L 85.234 20 L 73.418 3.252 L 73.262 3.252 L 73.262 20 L 71.436 20 L 71.436 0 L 73.203 0 L 85.06 16.768 L 85.216 16.768 L 85.216 0 L 87.013 0 Z M 53.418 20 L 51.504 20 L 58.78 0 L 60.733 0 L 68.008 20 L 66.094 20 L 59.824 2.373 L 59.688 2.373 Z M 37.09 0 L 43.34 17.627 L 43.516 17.627 L 49.766 0 L 51.68 0 L 44.404 20 L 42.451 20 L 35.176 0 Z M 20.923 20 L 19.009 20 L 26.284 0 L 28.238 0 L 35.513 20 L 33.599 20 L 27.329 2.373 L 27.193 2.373 Z M 15.576 0 L 15.576 20 L 13.8 20 L 1.982 3.252 L 1.826 3.252 L 1.826 20 L 0 20 L 0 0 L 1.768 0 L 13.623 16.768 L 13.779 16.768 L 13.779 0 Z M 27.078 18.156 C 28.502 18.156 29.656 17.002 29.656 15.578 C 29.656 14.154 28.502 13 27.078 13 C 25.654 13 24.5 14.154 24.5 15.578 C 24.5 17.003 25.654 18.157 27.078 18.157 Z M 59.578 18.657 C 61.002 18.657 62.157 17.503 62.157 16.078 C 62.157 14.654 61.002 13.5 59.578 13.5 C 58.154 13.5 57 14.654 57 16.078 C 57 17.502 58.154 18.656 59.578 18.656 Z M 99.078 18.156 C 100.011 18.174 100.881 17.687 101.353 16.882 C 101.825 16.077 101.825 15.079 101.353 14.274 C 100.881 13.469 100.011 12.982 99.078 13.001 C 97.654 13.001 96.5 14.155 96.5 15.579 C 96.5 17.002 97.654 18.157 99.078 18.157 Z" fill="transparent" height="20px" id="jyq5QjDNY" width="106.94799969482423px"><path d="M 92.359 20 L 90.445 20 L 97.72 0 L 99.673 0 L 106.948 20 L 105.034 20 L 98.764 2.373 L 98.628 2.373 L 92.358 20 Z M 87.012 0 L 87.012 20 L 85.234 20 L 73.418 3.252 L 73.262 3.252 L 73.262 20 L 71.436 20 L 71.436 0 L 73.203 0 L 85.06 16.768 L 85.216 16.768 L 85.216 0 L 87.013 0 Z M 53.418 20 L 51.504 20 L 58.78 0 L 60.733 0 L 68.008 20 L 66.094 20 L 59.824 2.373 L 59.688 2.373 Z M 37.09 0 L 43.34 17.627 L 43.516 17.627 L 49.766 0 L 51.68 0 L 44.404 20 L 42.451 20 L 35.176 0 Z M 20.923 20 L 19.009 20 L 26.284 0 L 28.238 0 L 35.513 20 L 33.599 20 L 27.329 2.373 L 27.193 2.373 Z M 15.576 0 L 15.576 20 L 13.8 20 L 1.982 3.252 L 1.826 3.252 L 1.826 20 L 0 20 L 0 0 L 1.768 0 L 13.623 16.768 L 13.779 16.768 L 13.779 0 Z" fill="rgb(255, 255, 255)" height="20px" id="p4NXWnxRX" width="106.94799969482423px"/><path d="M 2.578 5.156 C 4.002 5.156 5.156 4.002 5.156 2.578 C 5.156 1.154 4.002 0 2.578 0 C 1.154 0 0 1.154 0 2.579 C 0 4.003 1.154 5.157 2.578 5.157 Z" fill="rgb(255, 109, 56)" height="5.157000000000011px" id="m55y8SavT" transform="translate(24.5 13)" width="5.156499999999994px"/><path d="M 2.578 5.157 C 4.002 5.157 5.157 4.003 5.157 2.579 C 5.157 1.154 4.002 0 2.578 0 C 1.154 0 0 1.154 0 2.578 C 0 4.002 1.154 5.156 2.578 5.156 Z" fill="rgb(161, 96, 96)" height="5.157000000000011px" id="rCTLhUac_" transform="translate(57 13.5)" width="5.156500000000051px"/><path d="M 2.578 5.156 C 3.511 5.174 4.381 4.687 4.853 3.882 C 5.325 3.077 5.325 2.079 4.853 1.274 C 4.381 0.469 3.511 -0.018 2.578 0.001 C 1.154 0.001 0 1.155 0 2.579 C 0 4.002 1.154 5.157 2.578 5.157 Z" fill="rgb(255, 190, 86)" height="5.156502122779082px" id="YsCNByWzb" transform="translate(96.5 13)" width="5.206771547149877px"/></g></svg>)
API PLATFORM
API PLATFORM
Build voice into your products with speech & voice APIs for India
Build voice into your products with speech & voice APIs for India
Transcribe and speak like India does with Bodhi STT and TTS APIs. Powered by foundational, compute efficient models built for enterprise scale.
Transcribe and speak like India does with Bodhi STT and TTS APIs.
Powered by foundational, compute efficient models built for enterprise scale.
The full-stack Voice API platform to build products that India can use
The full-stack Voice API platform to build products that India can use
LISTEN
LISTEN
Speech to Text
Speech to Text
Real-time and batch transcription built for Indian languages, accents and code-mixed speech.
Real-time and batch transcription built for Indian languages, accents and code-mixed speech.
SPEAK
SPEAK
Text to Speech
Text to Speech
Natural sounding, low-latency voices for multilingual conversational experiences.
Natural sounding, low-latency voices for multilingual conversational experiences.
PERSONALISE
PERSONALISE
Voice Cloning
Voice Cloning
Create custom voices from recorded or uploaded reference audio while retaining character and vocal identity.
Create custom voices from recorded or uploaded reference audio while retaining character and vocal identity.
UNDERSTAND
UNDERSTAND
Keyword Boost
Keyword Boost
Improve transcription accuracy for domain-specific words, names and terminology using custom hotwords.
Improve transcription accuracy for domain-specific words, names and terminology using custom hotwords.
CONTROL
CONTROL
Pronunciation
Pronunciation
Control how names, brands and domain-specific terms are spoken in your language.
Control how names, brands and domain-specific terms are spoken in your language.
ANALYSE
ANALYSE
Audio Intelligence
Audio Intelligence
Turn conversations into structured intelligence by extracting signals, entities, topics, etc from audio.
Turn conversations into structured intelligence by extracting signals, entities, topics, etc from audio.
Powered by flagship models built in-house
Powered by flagship models built in-house
Bodhi STT
Bodhi STT
Low-footprint speech recognition built to run efficiently on CPU, with deep coverage across Indian languages and dialects.
Low-footprint speech recognition built to run efficiently on CPU, with deep coverage across Indian languages and dialects.
Bodhi TTS
Bodhi TTS
Low-latency speech generation with 50+ voices across various Indian languages, curated for how India actually speaks.
Low-latency speech generation with 50+ voices across various Indian languages, curated for how India actually speaks.
Bodhi SLM
Bodhi SLM
A compact language model trained on BFSI conversations, designed to run on-prem alongside the rest of the Navana stack.
A compact language model trained on BFSI conversations, designed to run on-prem alongside the rest of the Navana stack.
Built for the way a billion people speak
Built for the way a billion people speak
Speech models built for India.
Fine-tuned for scale.
Trained on live field data, built for real-world noise, and hardened through millions of production calls.
Modular capability that lets you fine-tune the model to fit your use-case.
Speech models built for India. Fine-tuned for scale.
Trained on live field data, built for real-world noise, and hardened through millions of production calls.
Modular capability that lets you fine-tune the model to fit your use-case.
हिन्दी
मराठी
தமிழ்
తెలుగు
ಕನ್ನಡ
ଓଡିଆ
മലയാളം
বাংলা
অসমীয়া
भोजपुरी
मैथिली
मगही
ગુજરાતી
ਪੰਜਾਬੀ
English
Say it Kaise bhi, we get it
Language isn’t simple. It’s emotional, regional, nuanced and full of surprises. Bodhi gets that.
Speech models don’t work off the shelf in production. Ours let you tweak, tune, and optimise for your exact domain, down to the last detail.
From Hinglish to dialects, from sentence fragments to mumbled phrases, our models adapt to how people actually talk.
Say it Kaise bhi, we get it
Language isn’t simple. It’s emotional, regional, nuanced and full of surprises. Bodhi gets that.
Speech models don’t work off the shelf in production. Ours let you tweak, tune, and optimise for your exact domain, down to the last detail.
From Hinglish to dialects, from sentence fragments to mumbled phrases, our models adapt to how people actually talk.
Built for production. Proven in the real world.
Built for production. Proven in the real world.
Navana’s voice stack is already powering high-volume, multilingual conversations across enterprise use cases - with models built specifically for Indian speech.
Navana’s voice stack is already powering high-volume, multilingual conversations across enterprise use cases with models built specifically for Indian speech.
100M+
100M+
Voice minutes processed
Voice minutes processed
12+
12+
Indian languages currently live
Indian languages currently live
500+
500+
AI agents live across BFSI
AI agents live across BFSI
The numbers speak for themselves.
The numbers speak for themselves.
On a public benchmark of 67,283 real speech clips across nine Indian languages, Bodhi STT achieved the lowest transcription error rate in every language tested.
On a public benchmark of 67,283 real speech clips across nine Indian languages, Bodhi STT achieved the lowest transcription error rate in every language tested.
67,283 CLIPS
67,283 CLIPS
9 LANGUAGES
9 LANGUAGES
6 PUBLIC DATASETS
6 PUBLIC DATASETS
5 PROVIDERS
5 PROVIDERS
INDEPENDENTLY SCORED
INDEPENDENTLY SCORED
BODHI
87%
SARVAM
83.3%
AZURE
82.8%
SMALLEST
78.8%
DEEPGRAM
73.4%
#1
#1
IN 9 INDIAN LANGUAGES
IN 9 INDIAN LANGUAGES
Lowest word error rate across every Indian language benchmarked.
37%
37%
FEWER ERRORS
FEWER ERRORS
Than the average competing speech provider across the nine-language benchmark. Up to 51% fewer vs the weakest, and 11–22% fewer than the closest rivals.
Than the average competing speech provider across the nine-language benchmark. Up to 51% fewer vs the weakest, and 11–22% fewer than the closest rivals.
77%
77%
CONVERSATION RETAINED MEANING
CONVERSATION RETAINED MEANING
Bodhi preserved the speaker’s intended meaning in 77.4% of clips. Accuracy isn't just about getting every word right. It's about understanding what was said.
Bodhi preserved the speaker’s intended meaning in 77.4% of clips. Accuracy isn't just about getting every word right. It's about understanding what was said.
Give your product a voice.
Give your product
a voice.
Unlock the true potential of your business with Navana’s Indic Voice API Stack
Unlock the true potential of your business with Navana’s Indic Voice API Stack
Navana.ai
Navana enables last mile access of digital services to end users through the use of voice-based technologies and products.
Nanavati Mahalaya
18, Homi Mody St. Fort,
Mumbai, Maharashtra 400001
Navana.ai
Navana enables last mile access of digital services to end users through the use of voice-based technologies and products.
Nanavati Mahalaya
18, Homi Mody St. Fort,
Mumbai, Maharashtra 400001