Service providers
Updated October 22, 2026
Input uses service providers for hosting, AI, search, payments, communications, and analytics. This page supplements our Privacy Statement. The features you use and their settings determine which providers process your information.
AI and search
These features process relevant message and attachment content, drafts, instructions, conversation context, and contact details. Indexing, organization, summaries, and writing assistance can run automatically.
Google Cloud, through Vertex AI: Gemini models for briefs, contact research with Google Search grounding, and other AI tasks using relevant content and contact context.
Amazon Web Services, through Amazon Bedrock: AI assistance using models including Anthropic’s Claude, hosted and operated by AWS. Requests can include relevant messages, instructions, and conversation context.
OpenAI: AI features and fallback models, semantic-search embeddings, and audio transcription. Embedding requests can include email subjects, message text, and extracted attachment text; transcription processes submitted audio.
Cerebras: Autocomplete and grammar checking using relevant draft text and context.
TypeSafe AI: Classification and safety checks using relevant message content, instructions, contact details, or proposed action details.
turbopuffer: Search storage and retrieval using message text, metadata, and numerical embeddings.
Choosing a model applies to the relevant feature. Other features may use separate providers, and fallback models may use another provider when a request fails. Your email provider does not determine all AI processing providers.
AI services can use temporary storage, caching, and safety or abuse monitoring. Provider retention depends on the service and applicable configuration. Input also stores conversation history and search indexes under the retention practices described in the Privacy Statement.
Infrastructure and operations
Amazon Web Services: Hosting, database, file storage, queues, and related infrastructure supporting accounts, mailbox content, attachments, and service operation.
Mixpanel: Usage events associated with account identifiers, names, email addresses, inbox and team information, IP addresses, and device information.
Stripe: Payments and subscriptions. Stripe handles payment credentials; Input receives billing and transaction information.
Resend: Service and invitation emails, including recipient details and message content.
Loops: Contact lists and product communications using contact and account information.
Our monitoring infrastructure uses Sentry software to process error and diagnostic information, which can include identifiers, request details, and error context.
Research, connected accounts, and other services
Contact research uses the AI services above and Google Search grounding, Exa for web search, and Firecrawl for website and image research. Relevant names, email addresses, feedback, and signature links may be sent to the services involved. Some profile-image lookups use Gravatar or GitHub. The Privacy Statement explains how research profiles are shared across inboxes.
Google and Microsoft provide connected email and calendar accounts and receive requests and updates needed to operate them. Optional integrations receive information needed for the actions you authorize; their own policies apply.
If notifications are enabled, Apple or your browser’s push service processes delivery information. Notifications can include sender details, subjects, or previews, depending on the notification and your settings.
Loading external images, generating previews, or following links can send requests to the host website. Input proxies supported email images and previews and removes certain tracking parameters.
For questions about providers or applicable safeguards, contact support@input.so.