Documentation Index
Fetch the complete documentation index at: https://docs.latitude.so/llms.txt
Use this file to discover all available pages before exploring further.
Data protection
Hosted Latitude is designed for teams that need AI observability while keeping production agent data within a defined European region. It stores and processes customer data and runs Latitude-managed inference in AWS eu-central-1 in Frankfurt, Germany. This covers data Latitude receives or creates while operating the platform, including traces, spans, sessions, annotations, scores, issues, evaluations, search indexes, and related metadata.European data and inference boundary
Latitude-managed inference includes the internal model calls that power product workflows such as:- issue discovery and clustering
- issue names and descriptions
- flagger review for supported failure categories
- evaluation generation and alignment workflows
- search, scoring, and reliability analysis
Tenant and project boundaries
Latitude scopes data by organization and project:- Organizations are the top-level access boundary.
- Projects scope traces, search, scores, issues, evaluations, and related workflows.
- Issue detection runs within a project, so teams can keep unrelated agents separate by routing them to separate projects.
Data minimization
Latitude reduces sensitive-data exposure throughout the telemetry pipeline:- SDKs redact common security-sensitive attributes before export by default.
- Customers can configure additional SDK redaction patterns for application-specific PII.
- Internal systems avoid exposing raw payloads when summaries, derived labels, scores, or references are enough.
- Customer-facing workflows use scoped project data instead of broad organization-wide processing.