# finic-ai/kairos

> Open-source real-time fraud and risk decisioning engine that approves, steps up, reviews or blocks sign-ups, logins, orders and payouts, keeping each decision with its signals. Tuned by coding agents over MCP, combining Finic's Grayson decision model with logistic regression and gradient-boosted trees.

- Magnitude: 3.3 out of 10 — Early signal
- Stars: 52 total · +33 stars measured 2026-11-03, 22:01–23:35 UTC
- Star trust: star growth looks organic
- Category: Security · Language: TypeScript · License: Apache-2.0 · Created: 2026-10-30 · Last push: 2026-11-03
- GitHub: https://github.com/finic-ai/kairos · Homepage: https://kairos.finic.ai · Page: https://gitnova.dev/en/r/finic-ai/kairos

## Useful for

- Block suspicious withdrawals in real time using device, card and address signals
- Run experiments comparing Grayson answers with gradient-boosted trees on your labels
- Let a coding agent build and publish a fraud workflow over MCP

## Why it’s here

- Star-counter measurements on 2026-11-03 (UTC), 22:01–23:35: 19 → 52 stars (+33). This is the change over that interval.
- The repository is 4 days old and already has 52 stars. With less than two weeks of history, there's no usual pace to compare the spike against yet.
- Hacker News: “Show HN: Kairos – Open-source risk and fraud decisioning platform” — 31 points, 2 h ago.

## Star trust

Star growth looks organic.

Star-trust labels are heuristics based on the repository’s behavior, not a check of every stargazer.

## Numbers

- Forks: 0
- Issues and pull requests: 13
- Watchers: 1
- Average over the last week: 8 per day
- Usual pace: too little history (under two weeks)
- Stars in the last hour (measured): 21

## Stars per day, last 5 days (oldest → newest, today is partial)

2026-10-30 … 2026-11-03: 0, 0, 3, 4, 33

## Hacker News

- Show HN: Kairos – Open-source risk and fraud decisioning platform — 31 points, 4 comments: https://news.ycombinator.com/item?id=50358314

## Spotted in now

- Spotted on Hacker News

## More in this category

1. **Verlintas/NOIR** — 7.4 · Breakout · Security · Kotlin · +455 stars measured 2026-11-03, 00:30–23:31 UTC
   A desktop CLI that audits Android privacy via adb: it lists apps with sensitive permissions, extracts reporting endpoints from APKs, scores risk, and proposes reversible fixes.
   Full card: https://gitnova.dev/en/r/Verlintas/NOIR.md
2. **anthropics/FlashProve** — 5.9 · Early signal · Security · Lean · +137 stars measured 2026-11-03, 15:31–23:31 UTC
   A Lean 4 research prototype for zero-knowledge proofs of AI inference: a provider proves which model produced a reply without revealing its weights. Not maintained.
   Full card: https://gitnova.dev/en/r/anthropics/FlashProve.md
3. **Gezine/CivSwitch** — 5.4 · Early signal · Security · JavaScript · +76 stars measured 2026-11-03, 00:30–23:31 UTC
   A toolchain for userland code execution on Nintendo Switch via an XSS in the Civilization 7 chat, using a Raspberry Pi Pico as a USB HID keyboard and a persistent savedata trigger.
   Full card: https://gitnova.dev/en/r/Gezine/CivSwitch.md
4. **yt1988/who-are-you** — 5.1 · Early signal · Security · HTML · +39 stars measured 2026-11-03, 07:38–23:31 UTC
   A privacy self-check page that runs 86 browser sniffers any site can run without permission, builds a portrait of your likely location, device and network, and flags contradictions that reveal a VPN or proxy.
   Full card: https://gitnova.dev/en/r/yt1988/who-are-you.md
5. **CircumferenceBanker/autodesk-inventor** — 5.0 · Early signal · Security · +100 stars measured 2026-11-03, 01:16–23:31 UTC
   The repository poses as an Autodesk Inventor 2026 installer, but actually it is a PowerShell script that downloads and executes code from an external domain under the guise of a "pre-activated" build. It looks like a malicious scheme for…
   Full card: https://gitnova.dev/en/r/CircumferenceBanker/autodesk-inventor.md

---

Magnitude (0–10) measures how fast and how unusually interest in a repository is growing right now. It is not a quality score.
Days are UTC. “So far today” is a fact; “expected by the end of the day” is a forecast.
Summaries and use cases are written by an LLM (DeepSeek V4.1 Flash) from the README and may be inaccurate: verify specific claims (benchmarks, speed, hardware) in the repository itself.
Data as of 2026-11-03 23:46 UTC, updated every 30 minutes.