01Input
Events & traits
Collect the facts
Send the product events and customer traits already flowing through your stack. Kite keeps every fact timestamped and replayable.
report.exported
{ format: "pdf" }
For SaaS founders and technical teams
Know where every customer is. Without querying five tools.
Customer data is scattered across product analytics, CRM, support, and billing. Kite turns those facts into one continuously computed customer state, so your team can act without writing queries or reconciling dashboards.
Join the waitlist · We review every application · No commitment
Works with
and any webhook.
and any webhook.The problem
Your customer exists in fragments.
Usage lives in product analytics. Conversations live in the CRM and support desk. Revenue lives in billing. To answer one question, your team writes a query, opens several dashboards, and still argues over which number is current.
The answer
One current state for every customer.
Kite does not replace your source tools. It interprets their facts through rules your team controls and continuously computes lifecycle, activation, health, journey progress, and risk. Ask where a customer is and get one explainable answer, not another dashboard to decode.
How it works
From fragmented facts to one answer.
Stop rebuilding customer context on demand. Kite turns existing events and traits into what is true now, continuously and for every account.
02Meaning
TypeScript definitions
Define what they mean
Encode lifecycle, journeys, health, and risk in your repo. Product, engineering, and CS review the same typed logic.
onboarding.complete
→ lifecycle.activated
03Output
Computed state
Act on one answer
Kite evaluates every customer continuously and exposes their current state directly or through the tools your team already uses.
acme_corp
activated · health 82
Customer Success as Infrastructure
Rules in TypeScript. History in Git. Deploys from your terminal.
Your Customer Success logic deserves the same rigor as your application code. With Kite, every rule is a typed function you can review in a pull request, validate in CI, compare between versions, and roll back through the CLI. No drag-and-drop builder is required.
states/lifecycle.tsTypeScript
// Define lifecycle from product behavior
import { defineLifecycle } from '@kitesdk/config'
export default defineLifecycle({
initialState: 'new',
states: {
new: {},
activated: {
enteredWhen: {
journey: 'onboarding',
status: 'complete',
},
},
at_risk: {
enteredWhen: { inactiveDays: 14 },
},
},
})$ kite init ✓ project created with lifecycle, journeys, health, rules, segments, and signals $ kite validate ✓ configuration is valid $ kite deploy --env test ✓ configuration deployed to test
Project structure
kite initkite.config.tsEvent and customer-trait schemas
states/lifecycle.tsLifecycle transitions
journeys/onboarding.tsOnboarding milestones
rules/activation.tsActivation conditions
segments/at-risk.tsAt-risk customer segments
health/overall.tsHealth derived from behavior
signals/engagement.tsEngagement indicators
One engineering workflow
- 01
kite init
Scaffold the model
- 02
pull request
Review every rule
- 03
kite validate
Catch invalid references
- 04
kite deploy
Ship an immutable version
test · kt_test_live · kt_live_
Configuration primitives
Six typed primitives. Every customer question has one right answer.
Not a generic label engine. Kite gives you purpose-built primitives — each answers a distinct question about your customers. Compose them in TypeScript and let the state engine keep every answer current.
Which one stage is the customer in?
Mutually exclusive stages like new, trial, activated, or cancelled. A customer starts in initialState and transitions when enteredWhen conditions match. Use it for durable stages, not temporary labels.
example definition
activated: { enteredWhen: { journey: "onboarding", status: "complete" } }
Inspect computed state from the CLI or dashboard. Every lifecycle transition is logged with its trigger event.
AI in Kite
An assistant for the questions your customer state creates.
Kite keeps state computation deterministic. AI works on top of that foundation to help you design, understand, and debug your Customer Success model without becoming the source of truth.
01
Model the system
Turn your value moment and churn risks into a reviewable starting point for lifecycle, journeys, health, segments, and activation rules.
kite ai model-success02
Close event gaps
Compare your configuration with its event coverage and generate a prioritized, stack-aware instrumentation plan.
kite ai generate-instrumentation-plan03
Explain every outcome
Diagnose a customer from computed state, event history, and rule logs, or identify each activation condition still missing.
kite ai debug-customer account_42Kite AI / diagnosisevidence grounded
$ kite ai why-not-activated account_42 Activation status: 2 of 3 conditions met [met] workspace.created [met] teammate.invited x3 [missing] integration.connected Evidence: no integration.connected event was found in the customer timeline. Suggested next step: verify instrumentation before changing the activation rule.
No invented dataNo automatic actions
You stay in control
Customer analysis requires explicit confirmation, known personal data is removed server-side, generated files are previewed before writing, and AI can be disabled entirely.
The platform
Built like infrastructure, because it is.
Kite runs as multi-tenant cloud infrastructure with API-first access, separate test and live environments, and a CLI that treats your Customer Success configuration like application code.
- 01One answer, with evidence
- Inspect complete customer state, journey progress, health components, event effects, and rule logs from the CLI or dashboard. Every answer traces back to the facts and rules that produced it.
- 02Recomputation, not migration
- Preview the aggregate impact of new definitions with a dry-run, then explicitly recompute one customer or all customers from retained raw events.
- 03Versioned like your code
- Keep authored TypeScript in Git, validate it in CI, compare deployed configuration versions, and activate an earlier version with the CLI.
- 04Webhooks for downstream tools
- Deliver derived changes to your own systems through HMAC-signed webhooks with retries, delivery history, testing, and redelivery.
Pricing
Start focused. Scale without friction.
Planned capacity for every stage, from your first state model to enterprise operations.
For technical founders
Starter
$79/ monthbilled annually
The essentials to encode your first customer success motion.
- 5 seats
- 1,000 MAU
- 5,000 playbook runs / month
- 3 active playbooks
For growing product teams
Pro
$159/ monthbilled annually
Make Kite the shared state layer for product and Customer Success.
- 20 seats
- 10,000 MAU
- 50,000 playbook runs / month
- Unlimited playbooks and webhooks
For scaled organizations
Enterprise
Custom
Custom capacity, security, and support for complex operations.
- Unlimited seats, MAU, and playbook runs
- Rule-based, explainable health scoring
- White-label and SSO / SAML
- Dedicated onboarding
Planned launch pricing. Tiers and limits may change before access opens.
FAQ
Questions, answered plainly.
What to expect before adding Kite to your stack.
01Who is Kite for?
Kite is built for SaaS founders and technical teams that already collect product events and want customer lifecycle, activation, health, and risk logic to be explicit, reviewable, and reliable.
02How long does setup take?
Setup depends on the event quality and the first states you choose to model. Teams start with one event source and one high-value workflow so the model can be validated before it expands.
03Do I need to replace Segment or PostHog?
No. Kite complements your analytics stack. Keep using Segment or PostHog for analytics and send the product events needed by your customer-state definitions to Kite.
04Is Kite a CRM?
No. Kite does not manage sales pipelines or CSM tasks. It computes customer state and sends changes through signed webhooks to your internal systems and the tools your team already uses.
05Do I need to instrument new events?
Not necessarily. You can reuse events you already track. Additional instrumentation is only needed when a state depends on product behavior you are not currently capturing.
06How are states prioritized?
Lifecycle transitions are evaluated in the order defined in your configuration. With the same ordered events, customer traits, and configuration version, Kite produces the same result.
07Can an account have more than one state?
An account has one current lifecycle state, but it can simultaneously belong to multiple segments, progress through multiple journeys, and have a health score.
08What happens when a rule changes?
The new rule becomes active after deployment. Existing accounts keep their current computed state until new activity is evaluated or you explicitly run a recomputation from retained events.
09Can I test changes before publishing them?
Yes. Run the engine locally, validate the complete configuration, deploy to the isolated test environment, compare versions, and preview historical changes with a recompute dry-run.
10How do I start using Kite?
Install the CLI, run kite init, then use kite validate and kite dev to test the generated project locally. Follow the documentation for setup guides, or explore the source code on GitHub.
11How much does Kite cost?
Starter starts at $79/month when billed annually, or $99 month-to-month. Pro starts at $159/month when billed annually, or $199 month-to-month, while Enterprise pricing is custom. These are planned launch prices and may change before access opens.
12What happens if Kite becomes unavailable?
The SDKs use timeouts and retries without blocking normal application execution after client construction. Delivery failures are reported through the configured error callback so your application can monitor or recover them.