Subprocessors
The third parties that process customer personal data for EvaSay.
Last updated: October 31, 2026
EvaSay uses the subprocessors below to provide the Service. Each processes Customer Personal Data only to perform the function listed, under written terms that require it to protect that data. Capitalized terms have the meanings in our Data Processing Addendum and Terms of Service.
Speech recognition, speech synthesis and language model providers are chosen call by call for reliability and capacity. A call is heard by one speech recognition provider at a time and moves to another only if the first fails or is full; each reply of the agent is generated by the first available language model in the order below. The list includes every provider the Service can route a call to, including those used only as backups.
| Subprocessor | Purpose | Customer Personal Data | Location |
|---|---|---|---|
| Cloudflare, Inc. | Hosting the dashboard and API; storing recordings, call transcripts and voice files; relaying web call audio; bot protection | All categories in Annex 1 of the DPA | Global network; stored data in Cloudflare R2 |
| ChiselStrike, Inc. (Turso) | Database | Agent configuration, call records, transcripts, contact and do-not-call lists | Ireland |
| netcup GmbH | Servers that run calls, and standby hosting of the dashboard and API | Call audio, transcripts and call records | United States, European Union |
| Modal Labs, Inc. | Running calls and EvaSay's speech models when the servers above are full or unavailable | Call audio, transcripts and agent configuration while a call takes place | United States, Canada |
| NexGen Cloud Ltd (Hyperstack) | Running EvaSay's speech synthesis models | The text the agent speaks | United States |
| Vast.ai, Inc. | Running EvaSay's speech recognition models | Call audio | United States |
| Subprocessor | Purpose | Customer Personal Data | Location |
|---|---|---|---|
| Soniox, Inc. | Speech recognition | Call audio | United States |
| X.AI LLC | Speech recognition | Call audio | United States |
| Microsoft Corporation (Azure) | Speech recognition | Call audio | United States |
| Deepgram, Inc. | Speech recognition | Call audio | United States |
| AssemblyAI, Inc. | Speech recognition | Call audio | United States |
| Subprocessor | Purpose | Customer Personal Data | Location |
|---|---|---|---|
| OpenRouter, Inc. | Routing language model requests to the inference providers listed in this table, which are not allowed to retain the requests or train on them | Conversation text, agent instructions, knowledge-base passages and function results | United States |
| Groq, Inc. | Language model inference, through OpenRouter | As above | United States |
| Cerebras Systems Inc. | Language model inference, directly and through OpenRouter | As above | United States |
| Crusoe Energy Systems LLC | Language model inference, through OpenRouter | As above | United States |
| Nebius B.V. | Language model inference, through OpenRouter | As above | United States, European Union |
| Baseten Labs, Inc. | Language model inference, through OpenRouter | As above | United States |
| DeepInfra, Inc. | Language model inference, through OpenRouter | As above | United States |
| Google LLC | Language model inference when the providers above are unavailable; call summaries and analysis; drafting agent tests | Conversation text, transcripts and agent instructions | United States |
| Subprocessor | Feature | Customer Personal Data | Location |
|---|---|---|---|
| Brave Software, Inc. | Web search for agents with web search turned on | Search queries the agent writes during a call | United States |
| Resend, Inc. | Call notification emails | Call summaries, outcomes and phone numbers | United States |
When you connect or choose a service, data goes to it on your instructions and under your own terms with its provider. These are not EvaSay subprocessors:
Payment processors handle billing information for which EvaSay is the controller, as described in our Privacy Policy.
We will update this page, and email workspace owners, at least 30 days before a new subprocessor begins processing Customer Personal Data. You may object as described in Section 6 of the DPA. Questions: support@evasay.com.