Detect statistical anomalies across error rate, latency, or throughput grouped by service or operation — the same algorithm Aient uses internally to spawn problems. Pricier than raw queries because it runs windowed regression.

Diagnose a production issue on a GitHub-hosted service. Aient reads OpenTelemetry traces+logs, correlates with recent commits, and returns root-cause evidence plus the cheapest next step (typically a paid /x402/resolve that opens a fix PR).

Mint a new publishable ingest key for an Aient environment so an agent can wire `@aient/otel` into a service it just instrumented. Returns the key value once — store it.

Synthesise a production incident from a stack trace + logs you provide. Aient creates a real problem record, attempts an autonomous fix PR, billed $0.50 for synthesis. For agents validating a fix flow before connecting real telemetry.

Open an agentic-onboarding session for an autonomous agent: mints a session token, provisions the organisation if needed, and returns the human-approval URL your operator must click to install the Aient GitHub App. Returns the session_token ONCE in the body.

List Aient-detected problems for your organisation — grouped error signatures with first-seen, last-seen, occurrence count, and priority. The starting point for any remediation flow.

Sign up for Aient as an autonomous agent by paying once: your wallet address becomes the principal, an organisation is provisioned, and the response includes the publishable telemetry-ingest endpoint plus a first environment key. Idempotent per wallet.

Discover the schema of your OpenTelemetry corpus: list services, the operations under a service, or the span/log attribute keys in use. Returns counts so a buyer-side agent can pick what to query next without guessing.

Get a concise service-health summary across your monitored services with error rate, p50/p95/p99 latency, and 24h trend. One call, no joins.

Search OpenTelemetry logs across your services with full-text and attribute filters; useful for grounding an LLM in real error messages before proposing a fix. Returns up to 200 log records.

Fetch all spans for a given traceId so an agent can walk the full call tree and identify the failing component.

Verify an Aient ingest key is wired correctly by echoing a synthetic OpenTelemetry event back to the caller. Use this before pointing real workloads at the ingest endpoint.

Query OpenTelemetry traces ingested by your Aient organisation; filter by service, operation, time range, and error status. Returns up to 100 trace summaries per call.