Tenant and application isolation
Every application belongs to one developer workspace. Agents, customers, calls, and recordings are scoped to that application and are not shared across tenants.
Security and privacy
Applications control the fields they send and which fields agents may see. Customer, agent, and developer access are scoped separately. This page describes the controls in the platform today and the ones still being built.
Every application belongs to one developer workspace. Agents, customers, calls, and recordings are scoped to that application and are not shared across tenants.
Customers, support agents, and developers authenticate through distinct flows with distinct authorization scopes. An agent cannot use a developer console, and a customer cannot enumerate agents.
Support tokens and media credentials are issued per call and expire quickly. They cannot be replayed to join an unrelated conversation.
Call audio uses WebRTC with peer-to-peer media where the network allows and a TURN relay fallback otherwise. Media is encrypted in transit on both paths.
Speaker tracks are transcribed on the agent's device. Raw audio is not sent to a third-party speech service as part of the default flow.
Recordings and transcripts are uploaded to private, S3-compatible storage using short-lived, call-scoped URLs, and are readable only by authorized reviewers.
We publish these as roadmap items rather than claims. Ask us for current status before making a compliance decision.
Recording consent
RingDesk provides consent-aware recording controls and records only when recording is enabled for the call. You remain responsible for the disclosure language, the legal basis, and the retention rules that apply in the markets you operate in.
Agent performance metrics are provided for operational visibility. They are not designed or marketed as an opaque employee-ranking system.
Scope limits are part of the security story.
Ask for the current architecture review before you build against RingDesk.