Trust & Security
How Ragently handles prompts, responses and metadata in AI processing.
AI processing through Anthropic (Claude) and OpenAI, with each as the automatic fallback for the other
Your customers' privacy is central to how Ragently is built. Ragently uses the business APIs of two AI providers: Anthropic (Claude) and OpenAI. For each function, one provider is the first choice and the other the automatic fallback. Neither uses the data to train AI models. Anthropic deletes API data within 30 days by default; OpenAI retains it for at most 30 days for abuse monitoring. The same applies to images uploaded in the chat.
To generate a response, Ragently sends the conversation and the relevant knowledge base context to the business API of one of two AI providers: Anthropic (Claude) or OpenAI. For each function, one provider is the first choice and the other the automatic fallback. Currently, for example, the AI phone agent runs on Anthropic (Claude), with OpenAI as fallback; the web chat and the other AI functions, including analysing uploaded images, run on OpenAI, with Anthropic (Claude) as fallback. This split may change per function; the arrangements on this page apply to both providers.
Your customer's browser sends a message to Ragently. Ragently forwards the request over HTTPS/TLS to the AI provider. The response follows the same encrypted path back.
Fallback route: if the provider that is the first choice for a function is temporarily unavailable, the other provider automatically takes over the request. With OpenAI, the storage option is switched off (store=false). See also sections 7, 8 and 10 of our privacy policy.
The best way to achieve privacy is to avoid collecting the information in the first place. Ragently stores conversation content for the inbox chat history and shares prompt and response content with AI providers solely to generate the response — never for training or permanent storage.
Anthropic deletes API inputs and outputs within 30 days by default; content flagged for a possible usage-policy violation can be kept longer for trust and safety (up to 2 years). OpenAI retains API data for at most 30 days for abuse monitoring.
Operational metadata: needed for authentication, billing, abuse prevention, reliability, and analytics. This is fully separate from actual prompt and response content.
Ragently uses the business APIs of Anthropic and OpenAI; for each function, one is the first choice and the other the automatic fallback.
Anthropic: our contracting party is Anthropic Ireland, Limited (6th Floor, South Bank House, Barrow Street, Dublin 4, D04 TR29, Ireland), part of Anthropic, PBC (United States). Processing and storage take place in the United States. Under Anthropic's Commercial Terms, Anthropic may not use customer content (inputs and outputs) from its business API to train models. Anthropic's Data Processing Addendum is part of those terms; under it Anthropic acts as processor, and as subprocessor for our customers. Transfers to the United States are covered by the Standard Contractual Clauses in that addendum.
OpenAI: Ragently uses only the business API, with the storage option switched off (store=false). OpenAI does not use API data for model training and retains it for at most 30 days for abuse monitoring.
The privacy and security documentation of both providers is available on their own websites.
We're happy to answer any questions about how Ragently protects customer conversations.
Get in touch