Join the Waitlist
The Contextual Speech API
for Voice Agents
Blynt handles the messy parts of Speech Recognition, so you can focus on your Agent's intelligence.
Built-in Natural Turn-Taking
Large scale Keyword Boosting
Run-time Contextualization
Interruption Handling
Context powers frontier LLMs.
It's time for Speech Models to be Context-Aware, too.
Meet the API that brings Context-Awareness to Speech AI
Built for Voice Agent Developers
Plug Blynt into your existing voice-agent framework
or build directly with our APIs and SDKs.
from livekit_plugin_blynt import BlyntSTT, BlyntSTTOptions, STTLanguages
# Create Blynt STT instance, and pass the options
blynt_stt = BlyntSTT(
api_key="your_api_key",
options=BlyntSTTOptions(
language=STTLanguages.EN,
session_entities=[
{ "name": "medical_term", "entity": "@acme_corp/medical_term" },
{ "name": "email", "entity": "@blynt/email" }
],
session_context=[
{ "name": "client", "value": "John Doe" },
{ "name": "pathologies", "values": ["hypertension", "asthma"] }
],
)
)
# Create the agent session, and pass the Blynt STT instance
session = AgentSession(
stt=blynt_stt,
llm="openai/gpt-4.1-mini",
tts="cartesia/sonic-3",
)
# Register the Blynt STT instance callbacks to the agent session
# so that it can access the conversational context and tts events
blynt_stt.register_callbacks(session)
await session.start(
room=ctx.room,
agent=Assistant()
)