Perplexity & DeepSeek Chat for VS Code

A personal extension I built for myself to seamlessly integrate the best AI models directly into my VS Code workflow. This tool combines the reasoning power of DeepSeek with the real-time search capabilities of Perplexity, allowing me to code without constant context switching.

✨ Key Features

• Multi-Model Support: Switch instantly between DeepSeek V3, DeepSeek R1 (Reasoner), Perplexity Sonar, and Sonar Pro.
• Dual API Integration: Connect both Perplexity and DeepSeek APIs for maximum flexibility.
• 📸 Vision & Image Analysis: Analyze images, screenshots, and diagrams (powered by Perplexity models).
• DeepSeek + Web Search: A unique feature that gives DeepSeek access to the web by using Perplexity as a search backend.
• Context-Aware: Attach your current file, specific code selections, or external text files to the chat context.
• Tabbed Interface: Manage multiple chat sessions simultaneously.

🛠️ Setup & Configuration

Since this is a personal tool, setup is straightforward but requires your own API keys.

1. Get API Keys

You will need API keys for the services you want to use:

• Perplexity API: Get it at perplexity.ai/settings/api
• DeepSeek API: Get it at platform.deepseek.com/api_keys

2. Connect in VS Code

1. Open the extension sidebar (click the Perplexity icon in the Activity Bar).
2. Click the "API" button in the chat interface.
3. Paste your keys into the respective fields:
   • Perplexity API Key: Required for Sonar models and Web Search features.
   • DeepSeek API Key: Required for DeepSeek V3 and R1 models.
4. Click Save and then Test to verify the connection.

🚀 How to Use

💬 Basic Chat & Model Selection

1. Click the "Model" button to choose your preferred AI.
   • Use DeepSeek R1 for complex logic and reasoning tasks.
   • Use Sonar/Sonar Pro for up-to-date information and web research.
2. Type your query and hit Enter.

📸 Search by Photo / Image Analysis

Note: Image analysis is currently supported only by Perplexity models (Sonar).

1. Click the Clip Icon (📎) next to the input field.
2. Select an image file (PNG, JPG, WEBP, etc.) from your computer.
3. The image will be attached to your message.
4. Type your question (e.g., "Explain this error screenshot" or "Convert this UI design to HTML/CSS").
5. Ensure a Perplexity model is selected, then send.

🌐 DeepSeek with Web Search

This extension enables a powerful hybrid mode: DeepSeek with Internet Access.

1. Select a DeepSeek model (V3 or R1).
2. Toggle the "Enable Search Mode" switch.
3. When you ask a question, the extension will first use Perplexity to search the web for real-time data.
4. It then feeds these search results to DeepSeek, allowing it to answer with up-to-date context.

📄 Managing Context

• Attach Current File: Press Ctrl+Alt+P (or Cmd+Alt+P on Mac) to add the currently open file to the chat.
• Add Selection: Select code in your editor and press Ctrl+Alt+S (or Cmd+Alt+S on Mac) to paste it into the chat.
• Context Toggle: Use the "Save Context" toggle to decide if the AI should remember previous messages in the conversation.

⚠️ Disclaimer

I created this extension for my own personal use to improve my productivity. It is provided "as is". Features may change as I tweak them for my workflow.

Enjoy coding with AI superpowers! 🚀