Prefactor
Observe, evaluate, and control AI agent runs in production
Find recurring AI agent failures and investigate production traces

Latitude is an open-source observability platform for teams running AI agents. It collects production traces through its SDK or an existing OpenTelemetry pipeline and analyzes completed conversations to identify recurring failures. Related traces become grouped issues with examples, trends, and affected users, helping developers prioritize patterns instead of reviewing every log individually.
Teams can define custom signals, annotate outputs, search sessions, and build versioned datasets from validated traces for evaluations or regression tests. Alerts can reach Slack, email, or webhooks. Latitude also connects to coding agents and provides an MCP interface, allowing investigation context to follow a developer into the workflow used to implement a fix.
Tried it? Share a useful thought or ask a question.
Good conversations start with a hello.
Sign in to comment