wustghj

wustghj / cursor-deepseek-v4-proxy

Public

解决 Cursor 使用 DeepSeek V4 模型时的 `reasoning_content must be passed back` 错误

18
0
69% credibility
Found May 02, 2026 at 18 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

This project sets up a personal bridge so Cursor users can chat with DeepSeek's advanced AI model, adding smart memory for thoughts and gentle speed limits for steady performance.

How It Works

1
📰 Discover the helper

You hear about a simple tool that lets you use a powerful AI called DeepSeek inside your Cursor coding app for better help.

2
📥 Get the files

Download the few easy files from the sharing site to your computer.

3
🚀 Start your bridge

Run the ready-to-go starter script, and it quickly sets up your personal connection to the AI.

4
🔗 Link to Cursor

Tell your Cursor app to chat through your new bridge by entering its web address.

5
💬 Begin chatting

Ask the AI questions in Cursor, and it remembers past thoughts and paces itself to stay smooth.

6
⭐ Smarter coding

Enjoy quick, clever suggestions that make your coding faster and more fun.

✅ Perfect setup

Your coding helper now uses the advanced AI reliably, saving you time every day.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 18 to 18 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is cursor-deepseek-v4-proxy?

This Python Flask proxy lets Cursor users tap into DeepSeek V4 models without the dreaded "reasoning_content must be passed back" error blocking their workflow. Point Cursor's DeepSeek API calls—like for cursor deepseek coder or cursor deepseek v3 agent—to localhost:9000/v1/chat/completions, and it patches message history, handles streaming responses, and proxies to api.deepseek.com. Fire it up with the included start script for local runs or Cloudflare tunnel exposure, fixing cursor deepseek 3.1 and cursor deepseek v3 free hiccups on the spot.

Why is it gaining traction?

It nails a hyper-specific pain point in cursor deepseek api and cursor deepseek hosting setups that official integrations ignore, with user-noticeable perks like automatic reasoning_content caching, rate limiting to dodge 429s, and retry logic for reliable streams. No more manual message fiddling or cursor github issues around bad credentials or mcp—developers get seamless cursor deepseek r1 or v3 2 chats right away. At 18 stars, it's word-of-mouth gold for Cursor power users tired of half-baked github copilot alternatives.

Who should use this?

Cursor devs grinding AI-assisted coding with deepseek v3 1 cursor or cursor deepseek v3 2, especially those hitting reasoning_content errors in multi-turn convos. Ideal for indie hackers prototyping cursor github codespaces or pull request reviews via DeepSeek, or anyone dodging cursor github token hassles. Skip if you're not deep in Cursor's ecosystem.

Verdict

Grab it if that exact error is killing your cursor deepseek flow—it's a quick win despite 18 stars and thin docs signaling early maturity. Credibility score of 0.699999988079071% means vet it locally first; solid for personal hacks, not enterprise yet.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.