This document is provided for transparency during our invite beta and is subject to revision.
Subprocessors
Last updated: June 12, 2026
Tally Technologies Inc. uses the following third-party subprocessors to provide the Tally service. Each subprocessor receives only the data necessary for its purpose.
| Provider | Purpose | Data shared | Location |
|---|---|---|---|
| AssemblyAI | Audio transcription (live and batch) | Customer audio and resulting transcripts | US |
| OpenAI | Text embeddings for search and retrieval | Transcript and content text chunks | US |
| Anthropic | AI analysis and insight generation | Transcripts, content excerpts, prep context | US |
| Amazon Web Services (incl. Bedrock) | Cloud hosting, database, file storage (S3), AI model inference | All customer content and account data | US |
| Perplexity | Web research for prep and live insights | Research queries derived from customer content | US |
| Stripe | Subscription payments and billing | Name, email, billing details, payment card (held by Stripe) | US |
| Resend | Transactional email delivery | Email address, name, notification content | US |
| Cloudflare | CDN, DNS, security, and edge hosting | IP addresses, request metadata, cached site assets | US |
| PostHog | Product analytics and feature flags | Usage events, organization identifiers, plan tier (no customer content) | US |
A Note on AI Providers
Customer content is sent to the AI providers above (AssemblyAI, OpenAI, Anthropic, AWS Bedrock, Perplexity) solely to provide the service — transcription, search, prep reports, and live-session insights. We use API tiers under which these providers do not train their models on our customers' data, where such tiers are offered.
Changes to This List
We will update this page before adding or replacing a subprocessor that processes customer content. Questions or objections: support@tally.fm.