Halt — TypeScript SDK

SaaS-aware, cross-language rate limiting. Per-user / per-API-key / per-plan limits, quotas, weighted endpoints, abuse controls, atomic Redis accuracy, and built-in observability — with a matching Python package.

📖 Full documentation: halt.afroawi.com

Install

npm install halt-rate
# optional: production store + metrics
npm install ioredis @opentelemetry/api

Quick start

import { RateLimiter, InMemoryStore, presets } from 'halt-rate';

const limiter = new RateLimiter({
  store: new InMemoryStore(), // use RedisStore in production
  policy: presets.PUBLIC_API, // 100 req/min per IP
});

const decision = await limiter.check(req);
if (!decision.allowed) {
  // respond 429, Retry-After: decision.retryAfter
}

Production Redis, Express/Next adapters, plan-based limits, quotas, penalties, and observability are covered in the docs → halt.afroawi.com/docs.

Features

• Algorithms: token bucket, fixed/sliding window, leaky bucket
• Keys: IP, user, API key, composite, or custom
• Atomic Redis store (Lua, cluster-safe, fail-open) + in-memory dev store
• Dynamic limits: change limits at runtime with PolicyRegistry / cached loaders — no restart
• SaaS: per-plan limits, quotas, weighted endpoints, abuse penalties
• Observability: StatsCollector + OpenTelemetry metrics
• Adapters: Express, Next.js, Hono, Fastify, GraphQL (Apollo)

Runtime support

Ships dual ESM + CJS with types. The core (limiter, algorithms, in-memory store) is edge-safe — runs on Cloudflare Workers, Deno, Bun, and Vercel Edge. RedisStore is Node-only (TCP); for edge distributed limits, inject any RedisClientLike (e.g. a fetch/REST client). Full guidance: halt.afroawi.com/docs.

Links

• Docs: https://halt.afroawi.com
• npm: https://www.npmjs.com/package/halt-rate
• Source & issues: https://github.com/surafel-kindu/halt

License

MIT