Subprocessors

Last updated August 20, 2026

Ambi Laboratories Inc. uses the third-party service providers below to operate Wink. Know exactly where your data goes and what each provider does with it.

This list is referenced by our Privacy Policy and Terms of Service.

Infrastructure & hosting

Supabase
Database, file storage, and backend services
United States
supabase.com
DigitalOcean
Cloud infrastructure, application hosting, and managed caching
United States
digitalocean.com
Vercel
Frontend application hosting and delivery
United States
vercel.com
Verda
GPU hosting for our self-hosted transcription models
Finland
verda.com
Amazon Web Services
Storage and compute for machine learning training datasets and model training
United States
aws.amazon.com
RunPod
GPU compute for machine learning model training
United States
runpod.io

Identity & communications

WorkOS
Authentication, single sign-on, and user identity management
United States
workos.com
Twilio
SMS and messaging delivery
United States
twilio.com

Speech & AI

Deepgram
Speech-to-text and audio processing
United States
deepgram.com
OpenAI
AI language model services
United States
openai.com
OpenRouter
Routing and delivery of AI language model requests, pinned to US-domiciled model providers with prompt collection and training disabled where the provider exposes that control
United States
openrouter.ai
Google AI Studio
AI language model services
United States
ai.google.dev
Anthropic
AI language model services
United States
anthropic.com
Groq
AI language model services
United States
groq.com
Tavily
Web search used by the in-product assistant
United States
tavily.com
Cohere
AI language model services
Canada
cohere.com
Pinecone
Embedding inference
United States
pinecone.io

Analytics & engineering

PostHog
Product analytics and feature flags
United States
posthog.com
SonarQube
Code quality and static security analysis
United States
sonarsource.com

Third-party subprocessors used to operate Wink, grouped by category
Name	Purpose	Location	Website
Infrastructure & hosting
Supabase	Database, file storage, and backend services	United States	supabase.com
DigitalOcean	Cloud infrastructure, application hosting, and managed caching	United States	digitalocean.com
Vercel	Frontend application hosting and delivery	United States	vercel.com
Verda	GPU hosting for our self-hosted transcription models	Finland	verda.com
Amazon Web Services	Storage and compute for machine learning training datasets and model training	United States	aws.amazon.com
RunPod	GPU compute for machine learning model training	United States	runpod.io
Identity & communications
WorkOS	Authentication, single sign-on, and user identity management	United States	workos.com
Twilio	SMS and messaging delivery	United States	twilio.com
Speech & AI
Deepgram	Speech-to-text and audio processing	United States	deepgram.com
OpenAI	AI language model services	United States	openai.com
OpenRouter	Routing and delivery of AI language model requests, pinned to US-domiciled model providers with prompt collection and training disabled where the provider exposes that control	United States	openrouter.ai
Google AI Studio	AI language model services	United States	ai.google.dev
Anthropic	AI language model services	United States	anthropic.com
Groq	AI language model services	United States	groq.com
Tavily	Web search used by the in-product assistant	United States	tavily.com
Cohere	AI language model services	Canada	cohere.com
Pinecone	Embedding inference	United States	pinecone.io
Analytics & engineering
PostHog	Product analytics and feature flags	United States	posthog.com
SonarQube	Code quality and static security analysis	United States	sonarsource.com

We update this page when we add or remove a subprocessor. Questions about this list can go to contact@trywink.io.