Tekedia · Jun 2026
Tekedia Capital publicly announced an investment in Kalpa Labs and described the company as building scalable generalist speech models for foundational AI and human-computer interaction.
Company research
Public company, workplace, funding, and market signals
Updated Jul 30, 2026
Overview
Kalpa Labs is a San Francisco-based YC F25 audio research lab building a generalist speech model and developer platform for conversational speech, with products for Studio, live Realtime calls, and an API.
Primary product
Conversational speech model platform (Studio, Realtime, API)
Founded
2025
Headquarters
San Francisco, United States
Team size
1-10
Industry
Technology, Information and Media
Sub-industry
Speech models / voice AI
Offices
San Francisco, United States
Business model
Developer API
Beta software product
B2B speech/voice platform
Funding
People & investors
Work & culture
Small, research-heavy team that emphasizes shipping the model, building the full stack in-house, and working with high autonomy on audio AI infrastructure.
Visa sponsorship
Limited
Compensation
Public job postings list compensation ranges of about $140K-$200K and $150K-$220K for San Francisco roles; no broader compensation policy is public.
Product & market
Pricing
Not publicly disclosed; public materials indicate beta access and a metered API model.
Differentiators
- —Generalist speech model unifying STT, TTS, voice cloning, and conversational audio tasks
- —Instruction following and in-context learning for speech
- —End-to-end training and in-house stack ownership
- —Low-latency / efficiency focus for real-time speech applications
- —Studio, Realtime, and developer API as a single product surface
Technology
Python
TypeScript
REST API
OpenAPI
npm package distribution
24 kHz PCM WAV audio
base64-encoded 16-bit PCM WAV
distributed training stack
Competitors
OpenAI
ElevenLabs
Deepgram
Microsoft Azure AI Speech
Google Cloud Speech/CCAI
AWS Transcribe / Polly / Connect
Scale & contact
Estimated revenue
≈$171K annual revenue (third-party estimate)
Estimated monthly visits
37
Traffic estimate as of Jun 2026
Latest news
Startuply.vc · May 2026
Profile article on Kalpa Labs' unified speech model strategy, model sizes, and the claim that its 800M parameter model cost under $1,000 to train.
Fondo · Feb 2026
Launch write-up summarizing Kalpa Labs' generalist speech model, product surface, and compute/pilot asks.
Analytics India Magazine · Oct 2025
Article on Kalpa Labs' speech-AI thesis, focusing on latency, token efficiency, and unified speech tasks.
Jobs and careers at Kalpa Labs
0 jobs at Kalpa Labs
No open positions
Check back later for new openings