Show HN: Kairos – Open-source risk and fraud decisioning platform

Kairos decides all sign-up, login, order, fee and payout as it happens, and keeps all decision alongside the signals rearward it. It is an open-source substitute to Sardine, Sift, Riskified and SEON, alongside two things they don't have: your coding delegate tunes it, and decision models capture fraud before you have the labels to train on.

• Real-time decisions: one call returns approve, stage up, assessment or block, alongside a hazard mark and the reasons, from the subject's history, the devices, cards and addresses it shares alongside others, and the session.
• Workflows: ask a decision model, mark alongside a model, branch, call webhooks, open cases, and decide again later.
• Queues and cases: assessment activity for people, whose resolutions rotate into labels.

• MCP server: Claude Code, Cursor or Codex peruse your decisions and build workflows, alongside all act recorded under its key.
• Experiments: your delegate proposes example variants, Kairos trains and tests them on your data, and the delegate ships the winner.

• Grayson: Finic's decision example answers questions concerning all petition in plain language. It plant alongside few labels and says unstructured data.
• Models: logistic regression and gradient-boosted trees (XGBoost-compatible) on your numbers and Grayson's answers, trained in Kairos or uploaded.

• SDKs for TypeScript (@finic-ai/kairos) and Python (finic-kairos), a CLI, an HTTP API, and self-hosting alongside Docker.

You dispatch Kairos what happens in your product: sign-ups, logins, orders, payouts, chargebacks. It decides all petition (approve, stage up, assessment or block) and keeps all decision alongside the signals rearward it.

What makes Kairos distinct is who tunes it. Connect Claude Code, Cursor, Codex or any delegate complete MCP, and the delegate can do the activity of a fraud ML engineer:

1. Build the workflow. It defines what to decide, which questions to ask a decision model, and which sectors to score.
2. Design experiments. It proposes variants: a decision example alone, logistic regression on your numbers, gradient-boosted trees, a stack of them, or the decision model's answers blended alongside your numbers in one model.
3. Let Kairos run them. The server trains all type on your oldest tagged data, chooses on the next, and tests on the newest, afterward reports ROC-AUC, PR-AUC, the acquire complete a baseline alongside its 95% interval, and precision and recall at your thresholds.
4. Pick the victor and container it. The delegate explains its choice, applies the type as the workflow's next version, and comes rear whenever new labels arrive.

You: "Our chargebacks went up. Find a improved example for shopify_order and publish it if it wins." Agent: designs six variants → create_workflow_experiment → says the results → "Logistic regression on the order's numbers affirmative Grayson's answers wins: ROC-AUC 0.949 on validation, 0.930 on test. Publishing it as v6." 

Kairos is built about Grayson, Finic's decision model. You ask it questions in plain tongue ("Does the shipping location appearance akin a reshipper?", "Is this email throwaway?") and it answers alongside calibrated probabilities. Decision models are powerful anywhere traditional ML is weak:

• Few labels. Grayson plant on day one, before you have the thousands of chargebacks a example needs to study from.
• Unstructured data. It says descriptions, emails, names, addresses, notes and item titles, which a gradient-boosted tree can't.
• Explanations. Every answer is a inquiry you wrote, so all decision says why.

Traditional models are powerful anywhere decision models are weak: they study your own outcomes precisely, from your own numbers, and they disbursal nothing per decision. Kairos lets you merge both: Grayson's answers rotate into features next to your sectors in a logistic regression or gradient-boosted trees (XGBoost-compatible, trained in Kairos itself), and experiments inform you which blend wins on your data.

Sign up at kairos.finic.ai. Create a sandbox environment, a concealed key, and you're prepared to dispatch events.

In the console, open Settings › MCP, rotate on the tool groups your delegate needs, and create a key. Then, in Claude Code:

claude mcp add --transport http kairos https://kairos.finic.ai/mcp --header "Authorization: Bearer ksk_…"

Cursor, Codex and VS Code activity the identical way: see MCP for agents. Then ask your delegate to appearance at your decisions, build a workflow, or run an experiment.

npm instal @finic-ai/kairos # or: pip instal finic-kairos

import { Kairos } from "@finic-ai/kairos"; const kairos = new Kairos({ apiKey: process.env.KAIROS_API_KEY!, baseUrl: "https://kairos.finic.ai" }); await kairos.defineAction("withdrawal", { subject: "user", hold: "Should this withdrawal be held since the document shows signs of fraud or takeover?", }); kairos.track("login", { userId: "user_42", deviceId: "dev_9", ipCountry: "GB" }); const decision = await kairos.decide({ type: "withdrawal", userId: "user_42", usd: 900 }); // { decision: "approve" | "step_up" | "review" | "block", riskScore, reasons, … }

You need Node.js 22, pnpm 10 and a Grayson API key from portal.finic.ai. No repository to set up: the server starts its own Postgres.

pnpm install cp .env.example .env # set GRAYSON_API_KEY pnpm kairos dev # the API on http://localhost:8787 and the console on http://localhost:5273

For your team, docker create up --build. See Self-hosting and Local setup.

pnpm kairos signup # the local server, a new account pnpm kairos login --profile prod --url https://kairos.finic.ai --email [email protected] pnpm kairos keys create "My server" # a concealed key for your code pnpm kairos actions define withdrawal --data @withdrawal.json pnpm kairos import events transactions.csv --type transaction --ts created_at --entity user=customer pnpm kairos decide withdrawal -e user=user_42 -p usd=900 pnpm kairos workflows run review_payout -i user=user_42 --wait pnpm kairos --help # all command

To have kairos on your path, run pnpm nexus --global in apps/cli.

• Shopify: decide all order, clasp risky ones for review, and study from chargebacks. To get the Kairos app for your store, attain the Finic squad on Discord or at [email protected].
• Inbound webhooks: obtain events from providers specified as Stripe. See Inbound webhooks.

The docs are at kairos.finic.ai/docs, and in docs/.

pnpm test # all tests; no network pnpm test:python # the tests of the Python SDK pnpm typecheck

• Discord: questions, ideas and show-and-tell, at discord.gg/4zRwRCK6Mx
• Email: [email protected]
• Contributing: issues and drag requests are welcome. Read Working on Kairos first.

Kairos is licensed under the Apache License 2.0.