Trace every call. Score every agent. Guard every output.
Open-source observability built for the agentic era. Native MCP tracing, real-time drift detection, and composable guardrails — in one SDK.
|The only platform that traces MCP tool calls as first-class spans. 4-span OTel hierarchy with W3C traceparent propagation across server boundaries.
One line to instrument Anthropic, OpenAI, Vercel AI SDK. Every call traced with latency, tokens, cost, and hallucination scores.
Flame chart visualization of your entire agent workflow. Click any span to inspect prompts, completions, and tool inputs/outputs.
Composite 0-1 metric combining hallucination rate, error rate, tool success rate, and latency health. Track it in CI. Alert when it drops.
EWMA + z-score anomaly detection on latency, cost, quality, and errors. Alerts you before dashboards do.
Detects internal contradictions in LLM outputs — negation, numeric, temporal, entity, and sentiment conflicts.
Chain safety checks declaratively: injection detection, PII redaction, topic filtering, output validation.
28 pattern categories across system overrides, role manipulation, encoding tricks, and jailbreak attempts. < 5ms per check.
Hash, mask, or remove sensitive data before it hits your logs. 7 PII patterns with Web Crypto SHA-256.
| Feature | Vigil | Langfuse | Arize | Helicone | LangSmith |
|---|---|---|---|---|---|
| MCP Tracing | |||||
| Agent Scoring | |||||
| Drift Detection | |||||
| Guardrail Chains | |||||
| Hallucination Detection | Paid | ||||
| Self-hosted Binary | Docker | Docker | |||
| License | MIT | MIT | Apache | Source-avail | Proprietary |