sigiro docs

sigiro is an observability backend for agents, not for dashboards. It reads OpenTelemetry. It answers the two questions that an automatic loop must answer before the loop repairs a fault:

1. What changed? — sigiro anomalies returns the shifts that deviate from the statistical baseline of the service. sigiro ranks these shifts. sigiro also groups them into incidents across all signals.
2. Why? — sigiro investigate returns one structured evidence block for each service. The block holds ranked findings, error breakdowns for each operation, LLM statistics for each model, sampled spans, log patterns, and a ready-to-run SQL query on every row.

Both commands are also endpoints: GET /v1/anomalies and POST /v1/investigate. An agent uses these endpoints to ask the same two questions. Refer to the API reference.

sigiro is one static binary. You do not run a collector. You learn no query language other than SQL.

Start here

• Quickstart — run the binary, send telemetry to it, and ask it a question. This takes five minutes.
• Instrument your code with an AI agent — give one prompt to an agent. The agent connects OpenTelemetry to sigiro for you.
• Hosted beta — host nothing yourself. Send OTLP to api.sigiro.com with a tenant key.

API reference

sigiro serves a document at /openapi.json. The reference comes from that same document. Therefore the reference always agrees with the API.

• API reference — five endpoints. Each endpoint has a request playground.

An agent that reads /openapi.json needs no other information from us.