---
title: "Magpie gateway: coding plan in one click"
description: "Add CodingPass to Magpie, the local model gateway, in one click: Sonnet 5.5, Opus 5.5 and GPT-6.1 for Claude Code, Codex and Pi, from $10/mo."
url: https://codingpass.ai/tools/magpie
language: en
updated: 2026-11-17
---
# Magpie, with a fixed monthly bill.

Import CodingPass into Magpie from the Keys page. One provider brings the Anthropic, Chat Completions and Responses endpoints, so Claude Code, Codex, OpenCode, Gemini CLI and Pi can all run Claude, GPT-6.1 and open-weight models on one monthly coding plan.

- Setup in 60s
- No surprise charges
- Full refund if you've used 5% or less

Magpie provider

```
Name        CodingPass
Anthropic   https://codingpass.ai
Chat        https://codingpass.ai/v1
Responses   https://codingpass.ai/v1
Key         cp-live-…
Models      GET /v1/models
```

Model ID for your plan

**Go**

claude-sonnet-5-5

**Plus**

claude-opus-5-5

**Pro**

claude-fable-5-2

## Any model in any coding agent.

Magpie is a free, MIT-licensed local gateway and menu-bar app for macOS, Windows and Linux, with tens of thousands of GitHub stars. Run Claude Code on Kimi, Codex on DeepSeek, or every agent on CodingPass. Just want to switch the provider of each agent, without a gateway? See [CC Switch](https://codingpass.ai/tools/cc-switch).

### Every agent's model in one place

Pick the model for Claude Code, Codex, Gemini CLI or OpenCode from one list in the menu bar. No config files to edit: Magpie changes the one setting that matters and keeps comments and formatting.

### Keeps going when a quota runs out

Add a second account or a routing group, and the gateway moves on to the next one when a quota runs out, so the agent never sees the error.

### Translates every API format

Anthropic Messages, OpenAI Chat and Responses, and Gemini, including streaming, tool calls and reasoning.

### Masks secrets, shows usage

Turn on masking in Settings → Privacy and keys or passwords in a prompt are swapped for placeholders before the request leaves your machine. Tokens, cache hits and cost per model and per session.

## Which plan for Magpie?

Every plan works with Magpie. Pick the Claude flagship you want as your default; GPT-6.1 Terra and all open-weight models come with every plan and share the same monthly allowance.

- [Full plan details](https://codingpass.ai/pricing#compare-plans)
- [Allowances for all 24 models](https://codingpass.ai/models)

### Go: Claude Sonnet 5.5

Best for Daily coding across several agents

$10/mo

$60 Sonnet allowance a month

**Requests / month**

≈2,050

**Per 5-hour window**

≈342req

**Model ID**

claude-sonnet-5-5

### Plus: Claude Opus 5.5

Best for Large refactors, long-running tasks

$30/mo

$200 Opus allowance a month

**Requests / month**

≈3,420

**Per 5-hour window**

≈570req

**Model ID**

claude-opus-5-5

### Pro: Claude Fable 5.2

Best for Several agents at once and the hardest problems

$100/mo

$500 Fable allowance a month

**Requests / month**

≈5,090

**Per 5-hour window**

≈848req

**Model ID**

claude-fable-5-2

Estimates from CodingPass traffic: about 100K tokens per request at roughly 5% input, 1% output and 94% cache read. One agent task usually makes several requests. Allowances are metered at official API prices.

## Three steps. About a minute.

Magpie keeps one CodingPass provider for every agent. Import it once from the Keys page, then pick its models in each agent.

1. ### Create a key
   Sign in with Google, GitHub or email and pick a plan. Then go to Dashboard → Keys → New key. Keys start with `cp-live-`.
2. ### Install Magpie 0.1.8 or later
   Download the app for macOS, Windows or Linux from usemagpie.ai, or install the CLI with the script. Import links need version 0.1.8 or later.
   `curl -fsSL https://usemagpie.ai/install.sh | sh`
3. ### Import from the Keys page
   Click Import to Magpie under any tab. Magpie shows the name, the hosts your prompts and key go to, and the models; nothing is saved until you click Add.
   `Keys → Import to Magpie`
4. ### Pick a CodingPass model
   In Magpie, choose a CodingPass model for each agent, then start a new session. [Get your request ID](https://codingpass.ai/tools#check).
   `codingpass/claude-sonnet-5-5`

acme-web Through Magpie zsh

```
$ codex
  model: codingpass/gpt-6.1-terra · via magpie

> Migrate the settings form to the new validation schema.

● Edited src/settings/form.ts (+58 −21)
● Ran pnpm test · 27 passing
```

GPT-6.1 Terra

$1.55 / $60.00 resets in 11d

## Questions about Magpie.

**Is this allowed? Could my account get banned?**

You use only your CodingPass key. You never log in to, share or relay anyone's account, so your own Claude or ChatGPT account isn't involved. Your requests follow each vendor's usage policy.

**Does usemagpie.ai see my key?**

No. The link puts everything after `#` in `https://usemagpie.ai/import#…`, and browsers do not send that part to the server. The key is stored only in Magpie on your computer. To change the key, import again: the link always uses the provider id `codingpass`, so Magpie offers Replace instead of adding a second provider.

**Do my requests go through a third party?**

No. Magpie's gateway runs on your machine and sends requests straight to CodingPass. Each agent uses the API it speaks natively, so Claude Code keeps prompt caching and thinking on the Anthropic endpoint.

**Which agents can use the CodingPass provider?**

Claude Code, Codex, OpenCode, Gemini CLI, Pi and Crush. Codex uses the Responses endpoint, which the import already includes.

**Which models does Magpie show?**

The ones your plan includes. The import does not list models: Magpie reads them from `/v1/models` with your key and picks up new ones on the next refresh. [Models on each plan](https://codingpass.ai/models).

**Nothing happened when I clicked Import.**

If Magpie is not installed, the page offers the download. If it is installed but older than 0.1.8, update it and click Import to Magpie again.

**What happens when I hit the 5-hour cap?**

Each 5-hour window allows 1/6 of the monthly allowance (on Go, $10 of Sonnet 5.5, about 342 requests). The cap counts all agents and models together, so CodingPass then returns 429 with the reset time for every agent. Wait for the reset, buy a reset card for the current window, or upgrade. [How caps and 429s work](https://codingpass.ai/usage-limits#response).

## Same key, every agent.
