---
title: "Write applications: a local broker and a client"
description: "Start one Queen node with Docker, open the dashboard, and add the client for JavaScript, Python, Go, Rust, PHP or C++. Every code block on these pages runs in all six."
---

> Queen MQ documentation, for AI agents
> Complete self-contained summary of Queen MQ: https://queenmq.com/llms-brief.txt
> Fetch that first when the question is about the product rather than about this page.
> Index of all pages: https://queenmq.com/llms.txt

# Write applications: a local broker and a client

Every feature of the client, one page each, with its reason. The code is in the six languages of the SDKs: pick yours in any code block, and every block on every page follows. The text names options as the JavaScript client spells them; [Client SDKs](/writing-applications/clients/) says how the others spell them.

## Start a broker

One binary serves the HTTP API and the dashboard on one port:

```bash
docker run -p 6632:6632 ghcr.io/queen-mq/queen:2.6.0
```

Open `http://localhost:6632` and watch every step of these pages land in the dashboard.

**Screenshot.** The Queen dashboard: the overview page of a running broker, with its queues, message rates and the replicated log.

The dashboard, served on the same port as the API.

**Figure.** One Queen process in a container. On port 6632 it serves the HTTP API that the SDKs and curl use, the dashboard in the browser, /health and the Prometheus metrics. Its data lives in one directory, /var/lib/queen/raft, which a volume keeps.

One process serves the API, the dashboard, health and metrics on port 6632.

## Add the client

| Language | Install |
| --- | --- |
| JavaScript | `npm install queen-mq` |
| Python | `pip install queen-mq` |
| Go | `go get github.com/smartpricing/queen/clients/client-go/v2` |
| Rust | `cargo add queen-mq serde_json` and `cargo add tokio --features rt-multi-thread,macros` |
| PHP | `composer require queen-mq/php-client` |
| C++ | Header only: `clients/client-cpp/queen_client.hpp`, with cpp-httplib and nlohmann/json |

No SDK is needed to try anything: the API is JSON over HTTP, and `curl` is a complete client.

## On these pages

| Page | What it covers |
| --- | --- |
| [Client SDKs](/writing-applications/clients/) | Install and connect, how each SDK names things, and what each one lacks. |
| [Queues](/writing-applications/queues/) | Create and configure a queue: every option, with its default and its reason. |
| [Push](/writing-applications/push/) | One partition per entity, an answer per message, buffering, binary forms. |
| [Pop and consume](/writing-applications/pop/) | Queue mode and consumer groups, where a group starts, batches, leases, conflation, seek, reads by offset. |
| [Acknowledge](/writing-applications/ack/) | completed, failed, retry and dlq, one by one or in a batch. |
| [Retries and DLQ](/writing-applications/retries/) | How a failed message comes back, and how to read and replay the dead-letter queue. |
| [Transactions](/writing-applications/transactions/) | Acks, pushes, KV and timers in one commit. |
| [KV, locks and timers](/writing-applications/kv-locks-timers/) | State in the broker, singletons with leased locks, messages scheduled for later. |
| [Ephemeral queues](/writing-applications/ephemeral/) | Request and reply in memory, outside the log. |
| [Streams](/writing-applications/streams/) | Windows, reduce and gates, with the source ack, the state and the output in one commit. |
| [AI coding agents](/writing-applications/ai-agents/) | Give your coding agent the rules and the examples through MCP. |

Source: https://queenmq.com/writing-applications/index.mdx