# Open source
# OpenTelemetry native
# 10 minutes to first issue
Automatic issue detection for AI agents
Trace your agent in production. Latitude finds the failures and writes the evals
Agents fail differently
They don't only crash. They hallucinate, lose context, call the wrong tool, and confidently return the wrong answer. You need more than logs to catch that.
Search through thousands of traces
Find the exact step where your agent went wrong. Filter by error type, model, user, time range.
Automatic failure clustering
Latitude groups similar failures into issues without you configuring anything. No rules, no regex.
Evals generated from real failures
Every discovered issue becomes a running eval. New traffic is tested against known failure modes automatically.
Monitor recurring failure modes effectively with issues
Automatic detection
Latitude creates potential issues from your traces automatically. Review and validate them.
Alerts
Get notified when a new issue is detected or an existing one escalates on Slack, email or using webhooks
Hallucinating policy details
Pending review
NSFW Speech
Monitoring
Regressing
Memory loss
Escalating
User frustration
Escalating
Monitor with evals
Create an eval from any issue. Latitude aligns it to your validated feedback and runs it against new traffic continuously.
Golden datasets
Latitude automatically builds a golden dataset for each issue from your validated traces.
Human signal
Latitude clusters your team's feedback into failure modes so nothing gets lost.
Get started in minutes
Set up Latitude in your project and discover first issues in as little as
$
npx -y @latitude-data/claude-code-telemetry install
What's the difference between LLM observability and regular logging?
Do I need observability if my LLM app is already working fine?
How is Latitude different from Langfuse or other observability tools?
How quickly can I see my first production traces?
What issues will observability actually help me catch?
We're already using OpenAI's dashboard. Why do I need more?
Once I can see issues, then what?
