Elite AI Empire · OurTools

The linter for AI output.

Deterministic, CI-native quality checks for every LLM & agent response — no LLM bill, no flaky judge, no proxy in your request path. Think ESLint, but for what your agents actually say.

See pricing How it works

Already running in report-mode across the Elite AI Empire agent fleet.

What it catches

Refusals / capability disclaimers

Catches "as an AI I cannot…" leaking to users

Leaked secrets

API keys, private keys, tokens (sk-, AKIA, ghp_, xox…)

PII exposure

Emails, SSNs, card-like numbers in output

Prompt-injection echo

Output parroting "ignore previous instructions"

Truncation

Mid-thought / cut-off responses

Placeholder text

Leftover TODO / lorem-ipsum / {{tokens}}

Invalid JSON

Claims to be JSON but does not parse

Empty output

Near-empty / degenerate responses

Drop-in

# CI: report-mode (never blocks)
cat agent_output.txt | agent-output-qa

# CI gate (fail the build on grade D/F)
IFIXAI_QA_ENFORCE=1 agent-output-qa agent_output.txt

# Inline (Python)
from empire_app_kit.qa import qa_report
qa_report(response_text, app="my-agent")   # logs a grade, never raises

Deterministic-first: ~80% of checks need zero model calls — fast, reproducible, effectively free to run. Optional LLM-judge escalation only for the ambiguous cases you opt into.

Pricing

Metered on checks/month, not per-seat. No surprise LLM bills — deterministic checks are near-free to run.

Free

$0 forever

✓Full deterministic ruleset (all checks)
✓CLI + pre-commit + GitHub Action
✓Report-mode, run locally
✓Up to 10k checks/mo on dashboard
✓Community rule packs

Start free

Pro

$29 /mo

✓100k checks/mo
✓Hosted dashboard + trend history
✓Custom rule packs
✓Slack + PR annotations
✓1 optional LLM-judge add-on
✓Email support

Start Pro

Team

$99 /mo

✓1M checks/mo
✓Unlimited repos/projects
✓RBAC + SSO-lite
✓PII + injection security pack
✓Config-as-code policy gates
✓Priority support
✓Overage $1 / 100k checks

Start Team

Enterprise

Let's talk

✓Self-host / VPC (no data egress)
✓SAML SSO + audit logs
✓Custom validators
✓On-prem PII
✓SLA + dedicated support
✓SOC2 / HIPAA path

Contact sales

Positioned against Guardrails AI, Arize/Phoenix, LangSmith, Braintrust, Helicone, Patronus, Lakera and Galileo — the deterministic, CI-native slot none of them own.