Framework Integrations

glide-mq ships official integration packages for popular Node.js frameworks. Together they cover HTTP APIs, real-time SSE streaming, DI-first application wiring, and an operator-friendly dashboard.

Available Integrations

Package	Framework	Pattern	Install
@glidemq/hono	Hono	Middleware + typed router	npm i @glidemq/hono
@glidemq/fastify	Fastify v5	Two-plugin registration	npm i @glidemq/fastify
@glidemq/nestjs	NestJS 10+	Module + decorators + DI	npm i @glidemq/nestjs
@glidemq/hapi	Hapi.js	Plugin + Joi validation	npm i @glidemq/hapi
@glidemq/dashboard	Express	Drop-in web UI	npm i @glidemq/dashboard

Common Features

The integrations share the same core queue model, but they expose it in different ways:

• Hono, Fastify, and Hapi expose the full HTTP queue API with SSE, schedulers, flow create/read/tree/delete, flow usage and budget endpoints, queue-wide usage summaries, and broadcast publish/SSE routes.
• NestJS exposes the same core APIs through decorators and DI instead of shipping an HTTP layer. For cross-language /flows access, use the core proxy or one of the HTTP wrappers.
• Dashboard focuses on monitoring and operations, including queue controls, flow usage and budget views, job streaming, and queue-wide usage summaries. It does not act as a general flow-creation API.
• Testing mode is available across the API wrappers and NestJS for in-memory development. Broadcast and summary routes still require a live connection.
• Graceful shutdown is built into each integration using the framework's lifecycle hooks or registry helpers.

Choosing an Integration

If you use...	Install
Hono (edge, Bun, Cloudflare Workers)	@glidemq/hono - type-safe RPC, edge-native
Fastify (high-performance Node.js)	@glidemq/fastify - encapsulation-aware, Zod validation
NestJS (enterprise, decorators, DI)	@glidemq/nestjs - @Processor, @InjectQueue, full lifecycle
Hapi (enterprise, Joi validation)	@glidemq/hapi - Joi schemas, access control, SSE
Express (dashboard UI)	@glidemq/dashboard - drop-in web dashboard, no build step
Express/Koa/other (API only)	Use the core HTTP Proxy - queue routes, flow create/read/tree/delete, queue events SSE, usage summary, and broadcast over HTTP

Quick Start

Every integration follows the same pattern - declare queues, register the plugin, get endpoints:

// Example with Hono
import { Hono } from 'hono';
import { glideMQ, glideMQApi } from '@glidemq/hono';

const app = new Hono();
const connection = { addresses: [{ host: 'localhost', port: 6379 }] };

app.use('/mq/*', glideMQ({
  connection,
  queues: { emails: { name: 'emails' } },
}));

app.route('/mq', glideMQApi());

export default app;

See each integration's page for framework-specific setup, configuration options, and examples.

AI Framework Integrations

glide-mq works with popular AI SDKs via its AI-native job primitives. No additional packages are required - use the standard Queue and Worker classes.

Framework	Pattern	Example
Vercel AI SDK	generateText/streamText inside a Worker, job.stream() for token output, job.reportUsage() for metrics	with-vercel-ai-sdk
LangChain	LangChain chains inside a Worker, job.reportUsage() from response metadata	with-langchain

Vercel AI SDK

import { generateText } from 'ai';
import { Worker } from 'glide-mq';

const worker = new Worker('inference', async (job) => {
  const result = await generateText({
    model: openai('gpt-5.4'),
    prompt: job.data.prompt,
  });

  await job.reportUsage({
    model: 'gpt-5.4',
    tokens: {
      input: result.usage.inputTokens,
      output: result.usage.outputTokens,
    },
  });

  return { content: result.text };
}, { connection });

LangChain

import { ChatOpenAI } from '@langchain/openai';
import { Worker } from 'glide-mq';

const worker = new Worker('langchain', async (job) => {
  const response = await llm.invoke(messages);
  const usage = response.response_metadata?.tokenUsage;

  await job.reportUsage({
    model: 'gpt-5.4',
    tokens: {
      input: usage?.promptTokens ?? 0,
      output: usage?.completionTokens ?? 0,
    },
  });

  return { output: String(response.content) };
}, { connection });

See the AI Pipeline Examples for complete, runnable examples with both frameworks.