Nudgey is a behavioral nudge engine that sits between your app and your notifications. It figures out whether to nudge, what to say, and when to send it — so every message earns its place.
How It Works
Your app sends events. Nudgey handles the rest — evaluating rules, picking the best content variant, and waiting for the right moment.
Step 1
Your app fires events — signups, purchases, inactivity, whatever matters. Nudgey ingests them and builds a living profile of each user.
Step 2
Rules fire, risk scores run, and Thompson Sampling picks the best content variant. If the smart move is silence, Nudgey stays quiet.
Step 3
The nudge ships via webhook at the right time — respecting quiet hours, rate limits, and user fatigue. Then the feedback loop learns from what happens next.
Developer-First
Track events with the Python SDK. Nudgey handles rules evaluation, variant selection, timing, and delivery — your app never has to think about it.
RESTful API with 19 endpoints. Webhook-based delivery with HMAC-SHA256 signatures. Full analytics built in.
from nudgey import NudgeyClient client = NudgeyClient(api_key="nk_live_...") # Track a user event — Nudgey takes it from here client.track( user_id="usr_8a3f", event_type="cart_abandoned", properties={ "cart_value": 142.50, "items": 3, "minutes_idle": 45 } ) # That's it. Nudgey evaluates your rules, # picks the best message variant, waits for # the right moment, and delivers via webhook.
Under the Hood
Everything you need to send fewer, better messages — without building the infrastructure yourself.
Every user sees the content variant most likely to resonate. The model balances exploration and exploitation automatically — no manual A/B test setup required.
Three dismissed nudges in a row? User gets a break. Open rate tanking? Sending pauses. The system protects your users from fatigue without you writing a single rule.
Set quiet windows like 10 PM to 7 AM and Nudgey respects them in every user's local timezone. Cross-midnight windows work out of the box.
Every nudge decision is explainable. Your dashboard shows why a nudge was sent, which variant won, and the model's confidence. No black boxes.
Every query is scoped by tenant. ML models are stored per-tenant. There is zero cross-tenant data leakage — by design, not by policy.
Report opens, dismissals, and conversions back to Nudgey. The model retrains weekly on real engagement data — it gets sharper the longer you use it.
Works Everywhere
Nudgey was born in fintech and generalized to work anywhere users need a well-timed push.
Spending alerts, savings nudges, bill reminders — timed to when users are most receptive.
Cart recovery, restock reminders, post-purchase check-ins that don't feel like spam.
Workout reminders, streak protection, habit formation — respecting rest days and burnout signals.
Course completion nudges, study reminders, engagement recovery — tuned to each learner's pace.
What We Believe
Every query is tenant-scoped. Every model is isolated. User data stays where it belongs — we designed for this from day one, not as an afterthought.
If the smartest thing to do is not send a message, we don't send it. The best nudge engines know when to be quiet.
You can see why a nudge was sent, which variant was chosen, and how confident the model was. Transparency builds trust.
Get Started
We're onboarding design partners now. Get early access and help shape the product.