Data retention policy
Our data retention policy is governed by the Master Service Agreement we have with our customers
Data archiving and removal policy
Amplitude deletes customer data at the end of the customer’s subscription to the Amplitude services as set forth in the Amplitude Data Processing Addendum. During the term of a customer’s subscription, customers can, at their discretion, delete data from Amplitude using the deletion tools made available
Data storage policy
Data storage is indefinite unless otherwise dictated by the retention policy or archival / removal policy.
Data hosting details
Cloud hosted on AWS
App/service has sub-processors
no
App/service uses large language models (LLM)
yes
LLM model(s) used
Amplitude AI uses third‑party foundation models provided by our AI subprocessors, including OpenAI models (e.g., GPT‑4 family) and, for certain AI Agent and roadmap features, Anthropic models via Amazon Bedrock and Google’s Vertex/Gemini services.
LLM retention settings
Any retention of data by third‑party LLM providers is limited to direct service operation. Persistent logs and observability data is stored in Amplitude-controlled systems following our standard data-retention policies.
LLM data tenancy policy
Amplitude operates a multi‑tenant SaaS platform with strong logical isolation between customers. No customer data is retained or reused by LLM providers for model training, any telemetry occurs exclusively within Amplitude‑controlled systems
LLM data residency policy
AI features send data to LLM providers which act as subprocessors with documented regional and residency commitments. No customer data is stored outside of those regions and LLM data storage is exclusively transient for service operation.