Integrate ProxyAI with JetBrains IDEA for Seamless AI‑Powered Coding
This guide walks you through installing the ProxyAI plugin in JetBrains IDEA, configuring default and DeepSeek‑V3 models, obtaining API keys, and using features like chat, FIM code completion, bug detection, and code explanation to boost development efficiency.
1. Install the ProxyAI Plugin
Ensure your IntelliJ IDEA version is at least 2024.1. Open IntelliJ IDEA → Settings → Plugins → Marketplace, search for ProxyAI, and click Install . Restart IDEA when prompted.
2. Default Configuration
ProxyAI replaces tools such as GitHub Copilot, AI Assistant, Codiumate, and other JetBrains plugins by connecting local or remote models directly to IDEA. The default configuration covers two functional blocks:
Chat – for conversational interactions with the model.
FIM (Fill‑In‑The‑Middle) – for code and text completion.
Each block supports a set of free and paid models. The free models listed below are sufficient for most users: o3-mini (OpenAI) – ✅ gpt-4o (OpenAI) – ✅ gpt-4o-mini (OpenAI) – ✅ claude-3.7-sonnet (Anthropic) – ✅ gemini-pro-2.5 (Google) – ✅ gemini-flash-2.0 (Google) – ✅ qwen-2.5-coder-32b (Fireworks) – ✅ llama-3.1-405b (Fireworks) – ✅ deepseek-r1 (Fireworks) – ✅ deepseek-v3 (Fireworks) – ✅
These models can be selected directly in the ProxyAI settings without additional registration.
3. Configure DeepSeek‑V3 (SiliconFlow) Model
For higher‑quality responses you may use the paid DeepSeek‑V3 model provided by SiliconFlow. Follow these steps:
Open IntelliJ IDEA → Settings → Tools → ProxyAI and choose Custom OpenAI .
Enter your SiliconFlow API key (obtain it from the ProxyAI website under User settings).
Set the chat endpoint URL to https://api.siliconflow.cn/v1/chat/completions.
In the body configuration, set model to deepseek-ai/DeepSeek-V3 (select the model from SiliconFlow’s model marketplace).
For FIM, adjust the endpoint to https://api.siliconflow.cn/v1/completions and keep the body field model pointing to the same DeepSeek‑V3 model. Note that the FIM request must use the model field rather than message, otherwise the API will reject the call.
4. Practical Tips
4.1 Explain
The Explain action generates a natural‑language explanation of the selected code. Configure it under
IntelliJ IDEA → Settings → Tools → ProxyAI → Prompts → Chat Actions → Explain. If the response is in Chinese, add a translation step in the prompt to obtain English output.
4.2 Find Bug
The Find Bug action scans the selected snippet for potential issues and offers fix suggestions. After reviewing the suggested changes, click Apply to insert them at the cursor position, then use Accept All to commit or Reject All to discard.
Other features such as refactoring and optimization follow the same workflow and can be explored similarly.
5. Quick Demo
After configuration, you can test ProxyAI by selecting a piece of code and invoking the Find Bug or Explain actions. The plugin will display suggestions in a popup; you can merge them with Apply and finalize with Accept All.
References
JetBrains ProxyAIplugin page: https://plugins.jetbrains.com/plugin/21056-proxy-ai
SiliconFlow FIM user guide: https://docs.siliconflow.cn/cn/userguide/guides/fim
How this landed with the community
Was this worth your time?
0 Comments
Thoughtful readers leave field notes, pushback, and hard-won operational detail here.
