# Fidelity

How we make sure each simulator matches the real API — unit tests, workflow tests, real-world usage.

Every [simulator](https://www.rystic.ai/docs/concepts/glossary.md#simulator) is checked against the real API it models. Three gates, in sequence.

1. **Unit tests.** Every endpoint probed against the live API, field by field.
2. **Workflow tests.** Multi-step journeys replayed end to end.
3. **Real-world usage.** Our own bots and pipelines run against the simulators every day.

## The cards

One card per simulator: current match rate, verified workflows and every endpoint covered.

| Simulator | Matches live | Probes matching live | Workflows verified |
|---|---|---|---|
| Slack | 96.7% | 521 of 539 | 204 of 210 |
| GitHub Pulls | 94.7% | 588 of 621 | 375 of 398 |
| Kalshi | 91.7% | 331 of 361 | 61 of 69 |
| PayPal | 98.2% | 108 of 110 | 45 of 47 |
| Lob | 100.0% | 32 of 32 | 15 of 15 |
| Okta | 98.6% | 214 of 217 | 107 of 110 |
| Linear | 100.0% | 110 of 110 | 29 of 29 |
| Stripe | 98.8% | 1803 of 1824 | 1164 of 1181 |
| Resend | 100.0% | 64 of 64 | 22 of 22 |

---

Source: https://www.rystic.ai/docs/fidelity · Markdown: https://www.rystic.ai/docs/fidelity.md