XPlanator

XPlanator is a VS Code extension that explains selected code directly inside the editor.

It is built for developers who want a focused explanation of the code in front of them: what it does, how it works, and what to remember before changing it. It supports OpenAI-compatible APIs, Gemini, Claude, and local Ollama.

Features

Explain a selected block of code from the editor context menu
Explain the current line, symbol, function, or class under the cursor
Open a dedicated explanation panel from the editor title bar
Choose the generated sections: summary, details, and takeaway
Request a simpler or more detailed follow-up explanation
Copy individual explanation sections
Use OpenAI-compatible APIs, Gemini, Claude, or local Ollama
Store API keys with VS Code SecretStorage
Configure model, language, timeout, and context limits
Optionally include nearby file context and related workspace snippets

Why Use It

XPlanator is not a code generator. It is a reading and review tool.

Use it when you are onboarding to a codebase, reviewing AI-generated code, learning a new language or framework, or trying to understand a local implementation before editing it. The extension keeps the workflow inside VS Code and limits the request to the code context you choose.

Privacy And Provider Model

XPlanator does not run a hosted backend.

When using Ollama, requests are sent to your configured local Ollama endpoint, usually http://localhost:11434.

When using a cloud provider, the selected code and configured context are sent directly from VS Code to that provider using your own API key. API keys are stored through VS Code SecretStorage and are not written to settings.json.

Security behavior:

API keys are not stored in plain text settings
API keys are not logged by the extension
Saved API keys can be deleted from the settings panel
XPlanator does not proxy requests through a third-party service

See SECURITY_MODEL.md for implementation details.

Supported Providers

Provider	Default model	Notes
OpenAI / Codex	`gpt-5-nano`	Default provider, requires an API key
Ollama	`qwen2.5-coder`	Local provider, no API key required
Gemini	`gemini-2.5-flash`	Requires an API key
Claude	`claude-sonnet-4-5`	Requires an API key

Model names are editable from the settings panel. If a provider changes model names, update the configured model value.

Installation

Install XPlanator from the VS Code Marketplace, then reload VS Code if requested.

For local development:

npm install
npm run compile

To use the default Ollama model locally:

ollama pull qwen2.5-coder

Usage

Explain A Selection

Select code in the editor.
Right-click the selection.
Run Explain Code: Explain Selection.

You can also use the default keybinding:

Ctrl+Alt+E

On macOS:

Cmd+Alt+E

Explain At Cursor

Place the cursor on a line, variable, function, or class.
Open the Command Palette.
Run Explain Code: Explain At Cursor.

Open The Panel

Click the XPlanator icon in the editor title bar, or run:

XPlanator: Open Panel

Configure A Provider

Open the XPlanator panel.
Click Parametres.
Choose Ollama local, OpenAI / Codex, Gemini, or Claude.
Enter the model name.
For cloud providers, paste your API key.
Optionally run Check connection.
Click Enregistrer.

Settings

XPlanator exposes the following VS Code settings:

explainCode.provider: ollama, openai, gemini, or anthropic
explainCode.ollamaUrl: local Ollama URL, default http://localhost:11434
explainCode.openAiBaseUrl: OpenAI-compatible base URL, default https://api.openai.com/v1
explainCode.anthropicBaseUrl: Anthropic-compatible Messages API base URL, default https://api.anthropic.com/v1
explainCode.model: model used for explanations
explainCode.explanationSections: generated sections, any combination of summary, details, and takeaway
explainCode.responseStyle: simple, technical, or beginner
explainCode.language: fr or en
explainCode.timeoutSeconds: provider request timeout
explainCode.includeRelatedContext: include related workspace snippets when available
explainCode.maxRelatedSnippets: maximum number of related snippets
explainCode.maxRelatedLinesPerSnippet: maximum lines per related snippet
explainCode.maxContextCharacters: total context character budget

Non-local provider URLs must use https. Local endpoints such as localhost and 127.0.0.1 may use http.

Commands

XPlanator: Open Panel
Explain Code: Explain Selection
Explain Code: Explain At Cursor
Explain Code: Explain More Simply
Explain Code: Explain In More Detail
Explain Code: Open Settings
Explain Code: Check AI Provider Connection

Development

Run the extension locally:

Open this repository in VS Code.
Run npm install.
Run npm run compile.
Press F5.
Use the Extension Development Host window to test commands.

Run tests:

npm test

Package the extension:

vsce package

Roadmap

Marketplace demo assets
Project-level explanation mode
Dependency and call-flow view
Learning mode with quiz questions
Explanation history
Markdown export
Cost and token usage display for cloud providers
Open VSX publishing

See ROADMAP.md for more details.

Contributing

Contributions are welcome. Useful areas include provider compatibility, prompt quality, UI polish, documentation, and tests around explanation parsing.

Read CONTRIBUTING.md before opening a pull request.

Security

Please do not report security issues in public GitHub issues. See SECURITY.md.

License

MIT. See LICENSE.