Vitals · The interoception sense

vitel

Graded observability a coding agent consumes to feel whether the running system is healthy — and refuse to claim done while vitals are failing.
$ pip install vitel
What it does

Telemetry → a verdict

Grade & report

A machine verdict (pass/warn/fail) with issues grounded in a Metric — name, window, observed vs threshold, burn-rate. Deterministic, no key needed.

Match an SLO

Grade telemetry against an error budget — must: error_rate < 1%, must: error_budget_remaining > 20%, should: p99(latency_ms) < 300ms.

Analyze & critique

analyze adds anomaly detection + an optional LLM critique. Egress happens only here; check stays offline by construction.

Watch over time

watch grades behaviour over a window — slow latency regression, resource leak, error spike, flatline/liveness — bounded by construction.

Quickstart

One verdict away

$ pip install vitel && vitel demo

vitel demo grades a synthetic degrading service, prints a FAIL report (error budget burning — no LLM key), then grades the fixed version and prints PASS.

One ecosystem

Eyes, ears, vitals & brain

All share one contract, agentsensory — one Verdict, one Report, one Handoff.