Proofread and translate text in VS Code

This extension offers Proofread text and Translate text commands in VS Code. It's a simple alternative to DeepL, Grammarly, and other similar tools.

Notable features:

• Uses Copilot, Ollama, or OpenAI for proofreading.
• Configurable language model and prompts (with good defaults).
• Supports many target languages (default is English).

⚠️ By default, this extension uses the Copilot API. It's free with a limit of 50 requests per month, or unlimited for paid Copilot subscribers. Unfortunately, Copilot only supports proofreading, not translation. To use the extension for translation, switch to Ollama or OpenAI (see details below).

Installation

Open the VS Code marketplace (Ctrl+Shift+X / Cmd+Shift+X), search for the extension by typing "proofread" or "nalgeon.proofread", and click Install. You're all set!

Commands

Run these commands from the VS Code command palette (Ctrl+Shift+P / Cmd+Shift+P).

Proofread: Proofread Text

Proofreads the selected text and prints the suggested changes below it.

Proofread: Translate Text

Translates the selected text and overwrites it with the translation. Only available if proofread.ai.vendor set to openai.

Proofread: Set OpenAI API Key

Sets the value of the OpenAI API key. Only required if proofread.ai.vendor set to openai.

Using Ollama

Ollama runs AI models locally on your machine. Here's how to set it up:

1. Download and install Ollama for your operating system.
2. Set the environment variables to use less memory:

OLLAMA_KEEP_ALIVE = 1h
OLLAMA_FLASH_ATTENTION = 1

3. Restart Ollama.
4. Download the AI model Gemma 2:

ollama pull gemma2:2b

5. Change the Proofread settings:

proofread.ai.vendor = ollama
proofread.ai.model = gemma2:2b

That's it!

Gemma 2 is a lightweight model that uses about 1GB of memory and works quickly without a GPU. For good results, send only a few paragraphs at a time for proofreading or translation.

For larger documents or improved results, try models like mistral or mistral-nemo.

Using OpenAI

OpenAI offers state-of-the-art AI models through an API. Here's how to switch to it:

1. Get an API key from the OpenAI Settings.
2. When you have the key, set it using the Proofread: Set OpenAI API Key command.
3. Change the Proofread setting proofread.ai.vendor to openai.

That's it!

⚠️ OpenAI is a paid service; they charge for API use.

Settings

proofread.ai.vendor

Name of the AI model provider. Must be either copilot, ollama, or openai. Default: copilot

proofread.ai.url

Custom URL of the AI API endpoint (leave empty to use the default URL).

proofread.ai.model

Name of the AI model to use. Default: gpt-4o

proofread.ai.temperature

Sampling temperature to use (between 0 and 2). Higher values make the output more random, while lower values make it more focused and predictable. Default: 0

proofread.ai.timeout

Timeout for AI API requests in seconds. Default: 30

proofread.language

The language to proofread or translate the text into. Default: English (US)

proofread.prompt.proofread

The prompt to send to the OpenAI API for proofreading.

proofread.prompt.translate

The prompt to send to the OpenAI API for translation.

License

Created by Anton Zhiyanov. Released under the MIT License.