# Limitr: The Usage Runtime for AI Products

> Limitr is an embedded runtime that enforces what every user and agent can do, how much they get, and what it costs them, giving you control over cost-to-serve and revenue capture.

This file contains links to documentation sections following the llmstxt.org standard.

## Table of Contents

- [Concepts](https://limitr.dev/docs/spec/concepts.md): Every term used across the Limitr docs, defined in one place. Reference pages link here rather than re-explaining vocabulary.
- [SDK Quick Reference](https://limitr.dev/docs/spec/quickreference.md): A complete index of every method on the Limitr class, with descriptions and links to full documentation.
- [Quickstart](https://limitr.dev/docs/spec/quickstart.md): Define a policy and make your first usage calls, all in about five minutes.
- [Welcome](https://limitr.dev/docs/spec/welcome.md): Welcome to Limitr - the usage runtime for AI products. See Limitr live in action, reference the full Limitr policy spec, and find all of the materi...
- [Persisting and Restoring Customer State](https://limitr.dev/docs/guides/customerstate.md): How to snapshot and restore customer state with the local engine, and what the workflow looks like in Cloud mode.
- [Mapping Your Pricing Model to a Policy](https://limitr.dev/docs/guides/mapping.md): How to translate your existing pricing model into Limitr policy concepts — credits, entitlements, plans — before you write a line of code.
- [Handling Plan Changes](https://limitr.dev/docs/guides/planchanges.md): Upgrades, downgrades, and mid-period plan switches — the right call sequence, meter reset decisions, and grant behavior.
- [Testing Your Policy](https://limitr.dev/docs/guides/testingpolicy.md): Validation as a CI gate, enforcement tests for limits, events, grants, and feature gates, and testing policy changes before shipping.
- [Writing Custom Alert Conditions](https://limitr.dev/docs/guides/advanced/alertconditions.md): Define precise alert conditions in Stof — syntax, event payload fields, patterns for common use cases, and when to use setNotifications vs addHandler.
- [Capabilities](https://limitr.dev/docs/guides/advanced/capabilities.md): Named, callable units of policy logic — define them, expose them as Claude tool definitions, call them directly, and restrict access by plan.
- [Embedded Event Handlers](https://limitr.dev/docs/guides/advanced/eventhandlers.md): Attribute-decorated functions, registering TypeScript functions callable from Stof with doc.lib(), and the App namespace convention.
- [Writing Policies in Stof](https://limitr.dev/docs/guides/advanced/stofpolicies.md): Why to use Stof over YAML, the syntax basics, what Stof gives you that YAML doesn't, and a complete policy example.
- [Notifications & Alerting](https://limitr.dev/docs/guides/cloud/notifications.md): Configure subscriptions, scope alerts with segments and event conditions, and route notifications to email or the SDK.
- [Cloud Quick Start](https://limitr.dev/docs/guides/cloud/quickstart.md): Connect to Limitr Cloud, understand what changes, and migrate from local enforcement in one step.
- [Managing Customer Segments](https://limitr.dev/docs/guides/cloud/segments.md): Define dynamic customer segments using Stof conditions, use them to scope notifications and pricing rules, and filter your dashboard view.
- [Homepage](https://limitr.dev/docs/marketing-overview.md): Limitr is an embedded usage runtime for AI products: it enforces what every user and agent can do, how much they get, and what it costs them, runni...
- [Live Demo](https://limitr.dev/docs/spec/monetize/demo.md): An interactive monetization demo of Limitr, running live in your browser.
- [Credit Exchange](https://limitr.dev/docs/spec/monetize/exchange.md): Credit exchange table for unit and currency conversions.
- [Overview](https://limitr.dev/docs/spec/monetize/overview.md): An overview of how to define Limitr policies for monetization.
- [Credits](https://limitr.dev/docs/spec/reference/credits.md): Credit types, pricing models, unit conversion, overhead cost, price fields, and SDK methods.
- [Customers](https://limitr.dev/docs/spec/reference/customers.md): Customer type, creating and reading customers, alternative IDs, refs, metadata, and subscription billing.
- [Enforcement](https://limitr.dev/docs/spec/reference/enforcement.md): The enforcement API — allow, increment, decrement, check, set, value, remaining, limit, events, and event handler registration.
- [Entitlements](https://limitr.dev/docs/spec/reference/entitlements.md): Entitlement and limit schema, enforcement modes, resets, org scope, customer overrides, and SDK methods.
- [Exchange](https://limitr.dev/docs/spec/reference/exchange.md): The exchange table — rune system, exchange pairs, grant strategy, credit conversion, and multi-step chains.
- [Margin](https://limitr.dev/docs/spec/reference/margin.md): customerMarginSnapshot, marginSnapshot, how margin is calculated, and what affects accuracy.
- [Notifications](https://limitr.dev/docs/spec/reference/notifications.md): Notification type, event matching, in-policy Stof functions, in-process handling, and Cloud routing.
- [Plans](https://limitr.dev/docs/spec/reference/plans.md): Plan structure, billing period, subscription entitlement, trial period, and SDK methods.
- [Policy](https://limitr.dev/docs/spec/reference/policy.md): Every field, type, constraint, and behavior in a Limitr policy document.
- [Topups & Grants](https://limitr.dev/docs/spec/reference/topups.md): Topup definitions, grant lifecycle, reset modes, rollover, and SDK methods.
- [How to Design a Pricing Model That's Flexible and Scalable for AI Products](https://limitr.dev/blog/design-a-flexible-pricing-model-for-ai.md): A practical playbook for AI startups to design a flexible and scalable pricing strategy.
- [AI Pricing Strategy Guide for 2026: How to Cut Engineering Costs and Drive Revenue Growth](https://limitr.dev/blog/ai-pricing-strategy-guide.md): A practical playbook for AI startups navigating pricing complexity, margin pressure, and the shift to hybrid monetization.
- [We Mean Runtime Literally](https://limitr.dev/blog/we-mean-runtime-literally.md): Everyone in usage-based pricing says enforcement has to happen at runtime. Limitr means runtime as a noun — an execution environment embedded in yo...
- [How to Control Overhead Cost for AI Products](https://limitr.dev/blog/how-to-control-ai-overhead-costs.md): Cost-to-deliver isn't enough. Here's how to measure margin-to-deliver per account, feature, and vendor — then control it with enforcement limits an...