Data

Tools

Indexes

Applications

Experiments

Awesome

An open API service indexing awesome lists of open source software.

https://github.com/moritzfl/llm-subscription-usage-intellij

Plugin for IntelliJ IDEA that shows usage quota for ChatGPT subscriptions
https://github.com/moritzfl/llm-subscription-usage-intellij

Last synced: 6 days ago
JSON representation

Plugin for IntelliJ IDEA that shows usage quota for ChatGPT subscriptions

Awesome Lists containing this project

README 

          
# LLM Subscription Usage



Track and use your LLM subscriptions directly in IntelliJ IDEA — in the status bar, a detailed popup, through IDE chat tools, local AI client MCP configs, and a local OpenAI-compatible proxy.



This plugin started as a simple OpenAI quota checker. It has grown into a multi-provider subscription companion and now includes capabilities beyond quota display, such as hosted web search/image-generation MCP tools and a Codex-backed local proxy. The goal is to help you get the most value from the subscriptions you already pay for.



**Supports:** OpenAI (ChatGPT), Cursor, GitHub Copilot, OpenCode Go, Ollama Cloud, Z.ai, MiniMax, and Kimi.



  

    

      

        LLM Subscription Usage on JetBrains Marketplace

      

      


      LLM Subscription Usage on JetBrains Marketplace

      


      Install the plugin for IntelliJ IDEA.

      


      


      

        JetBrains Marketplace version

      

      

        JetBrains Marketplace downloads

      

      

        JetBrains Marketplace rating

      

    

  



---



## Features



**Status Bar Widget** — A compact indicator in your status bar shows at-a-glance quota status. Tooltip reveals quick details; click to open the full popup.



**Quota Popup** — Click the status widget to see:

- All your active subscriptions side-by-side

- Primary and secondary usage windows

- Next reset times

- Last refresh timestamps



**MCP Integration** — Exposes quota data and supported hosted capabilities to IntelliJ's built-in chat via the Model Context Protocol. It can also sync the currently running IntelliJ MCP server URL into JSON/TOML/YAML config files for local AI clients.



Individual MCP tools can be enabled or disabled in the IDE settings under `Tools` > `MCP Server` > `Exposed Tools`.



**MCP Web Search** — Search the web through subscription-backed provider APIs without leaving the IDE. Supported search providers:

- OpenAI/Codex via the existing OpenAI login

- Kimi via the existing Kimi login

- Z.ai via API key

- MiniMax via API key

- Ollama via API key



OpenAI/Codex search supports context-size, live-access, domain-filter, and optional source metadata controls.



**MCP Image Generation** — Generate images through hosted Codex tooling and optionally save the result directly to a file.



**OpenAI-Compatible Proxy** — Serves a local OpenAI-compatible API backed by your Codex subscription, so tools like JetBrains Junie can use it as a custom LLM provider.



**Customizable Display** — Drag-and-drop to reorder providers in the popup. Choose whether the indicator lives in the status bar or main toolbar.



**Automatic Refresh** — Quotas refresh every 5 minutes in the background, plus on login and when opening the popup.



**Secure Credential Storage** — OAuth tokens, API keys, session cookies, and device-flow credentials are stored in IntelliJ Password Safe.



---



## Installation



Open IntelliJ IDEA `Settings` > `Plugins` > `Marketplace`, search for **LLM Subscription Usage**, and click Install.



Or download a release ZIP from the [GitHub releases page](https://github.com/moritzfl/openai-usage-quota-intellij/releases) and install from `Settings` > `Plugins` > gear icon > `Install Plugin from Disk...`.



## Getting Started



1. Open `Settings` > `LLM Subscription Usage`

2. Login or add your credentials for your LLM Providers

3. Return to IDE — the status bar widget shows your quota

4. Click the widget for a detailed popup

5. Optional: enable `Sync IntelliJ MCP server URL to JSON/TOML/YAML files` to keep local AI client MCP configs pointed at IntelliJ's current MCP endpoint



---



## MCP Server URL Sync



The plugin can keep JSON, TOML, and YAML config files up to date with IntelliJ's current MCP server URL. Enable `Sync IntelliJ MCP server URL to JSON/TOML/YAML files` in settings, choose one or more config files, and select an existing string property to update.



This is useful because IntelliJ's MCP server port can change. Tools like OpenCode, Codex, or other local AI clients often store the IntelliJ MCP server URL in their own config files. Server URL sync keeps those configs pointed at the active IntelliJ MCP endpoint automatically, so they continue to connect after the port changes.



The settings page shows whether IntelliJ's MCP server is currently running, installed but stopped, disabled, or unavailable.



---



## OpenAI-Compatible Proxy



When you are logged in to OpenAI, the plugin can run a local proxy that exposes your Codex subscription through standard OpenAI-compatible endpoints (`/v1/chat/completions`, `/v1/responses`, `/v1/models`, plus LiteLLM-style `/v1/model/info`). Any tool that speaks the OpenAI API or expects a LiteLLM server can then use your subscription.



To enable it, open the **OpenAI** tab in settings, tick `Enable local OpenAI-compatible proxy`, and apply. The status line shows whether the proxy is off, starting, running, or failed. Use `Copy Base URL` and `Copy API Key` to configure clients; requests authenticate against the locally generated API key, which is stored in the IDE Password Safe.



Notes:



- Configure clients with the base URL **without** a `/v1` suffix (e.g. `http://127.0.0.1:14621`); clients append `/v1/...` themselves, and all routes also answer unprefixed.

- For JetBrains Junie, add the proxy as a LiteLLM provider with that base URL and the copied API key — the available models are then discovered automatically.

- Token refresh is owned by the plugin's regular OpenAI login; the proxy never stores credentials itself. If the upstream rejects a token, the proxy triggers a refresh through the IDE login flow and retries once.

- `Log requests and responses to disk` writes full request/response bodies to a temp folder for debugging. It is off by default, and logs are pruned automatically (7 days / 2000 files).



---



## Screenshots



### Quota Popup



![Quota popup](docs/quota-popup.png)



### Status Bar



![Status bar icon](docs/quota-statusbar-icon.png)



![Status bar with percentage](docs/quota-statusbar-percentage.png)



![Status bar with cake diagram](docs/quota-statusbar-cake.png)



### Chat (MCP)



![MCP integration](docs/quota-mcp-integration.png)



### MCP Server URL Sync



![MCP server sync settings](docs/mcp-server-sync.png)



### Settings



![Settings](docs/quota-settings.png)



---



## How It Works



The plugin calls each provider's usage API (with your OAuth token, API Key or other credentials) to fetch quota data, then displays it in a normalized format. Refresh happens automatically in the background every 5 minutes, plus on login and when opening the popup.



Quota data is stored locally in IntelliJ's secure credential storage. Raw responses as they arrived from the API endpoint are available in settings for transparency and debugging.



---



## Troubleshooting



**"Port 1455 is already in use"** — Another process is using the OAuth callback port. Stop it and retry login.



**"Not logged in"** — Open plugin settings and start the login flow again.



**Quota fetch errors** — If backend behavior changed, inspect `Last quota response (JSON)` in settings to see what changed.