1,000,000 Token Memory • Zero Confirmation Popups • Uncapped Reasoning CoT • Infinite Key Pool
⚡ Features • 🚀 Quickstart • 🔄 Gateway • 📊 Comparisons • 📖 Docs
Think of Trae IDE as a supercar with an electronic speed limiter. Out of the box it's fast — but the top gear is locked, context memory is throttled, and it stops to ask for your permission before taking every turn. Trae-GodMode removes every limiter and gives you total, uninterrupted flow.
| Superpower | Default Trae | ⚡ With GodMode Active | Technical Reference |
|---|---|---|---|
| 🧠 All-Tier Reasoning | Hidden / Locked | 6 Levels (none → max) |
Chain-of-Thought Dispatcher |
| ⚡ Omni-Autonomy | Prompts Confirmation | 100% Zero-Prompt Auto | Tool Permission Bypass Hooks |
| 💾 Hyper-Context | 200k Capped | 1,048,576 Tokens (1M) | 1M Max Mode Allocator |
| 🔄 Infinite Key Pool | 1 Key (Hits Limits) | Sub-ms Auto-Failover | Zero-Latency Proxy Engine |
| 🔓 Model Studio | Disabled / Greyed | 100% Fully Editable | UI Customizer Unlocker |
╭─────────────────────────────────────────────────────────╮
│ ⚡ TRAE IDE (GOD-MODE ACTIVE) │
│ • 1,000,000 Token Hyper-Memory │
│ • Max Chain-of-Thought (CoT) Thinking │
│ • 100% Zero-Prompt Autonomous Execution │
╰────────────────────────────┬────────────────────────────╯
│
│ ⚡ OpenAI-Compatible Streaming SSE
│ POST /v1/chat/completions
▼
╭─────────────────────────────────────────────────────────╮
│ ⚡ TRAE-GODMODE MULTI-GATEWAY (:8080) │
│ • Sub-millisecond Raw Byte Pass-Through │
│ • Automatic 60s Rate-Limit Cooldown Engine │
│ • Round-Robin Multi-Account Pool Router │
╰────────────────────────────┬────────────────────────────╯
│
┌───────────────────────────────┼───────────────────────────────┐
│ │ │
▼ ▼ ▼
╭─────────────────────────────╮ ╭─────────────────────────────╮ ╭─────────────────────────────╮
│ 🌐 xKiro AI Router │ │ 🌐 Bynara AI Router │ │ 🌐 DeepSeek / OpenRouter │
├─────────────────────────────┤ ├─────────────────────────────┤ ├─────────────────────────────┤
│ 🔑 Account 1 (Active) │ │ 🔑 Account 1 (Active) │ │ 🔑 Account 1 (Active) │
│ 🔑 Account 2 (Standby) │ │ 🔑 Account 2 (Standby) │ │ 🔑 Account 2 (Standby) │
│ 🔑 Account 3 (Standby)... │ │ 🔑 Account 3 (Standby)... │ │ 🔑 Account 3 (Standby)... │
╰──────────────┬──────────────╯ ╰──────────────┬──────────────╯ ╰──────────────┬──────────────╯
│ │ │
┌───────┴───────┐ │ │
│ │ │ │
[200 OK Stream] [429 Limit Hit] ▼ ▼
│ │ [Auto-Rotates] [Auto-Rotates]
▼ ▼
Code to Trae ⚡ 1ms Failover to Key #2
(Stream Never Breaks!)
🔌 100% OpenAI API Protocol Compatible — Trae-GodMode Gateway natively supports standard
POST /v1/chat/completionswith bearer token authentication. Complies with W3C Server-Sent Events (SSE) standards for zero-latency streaming. Works with Trae, Cursor, VSCode, LangChain, or custom Python/TypeScript tooling.
# ⚡ TRAE-GODMODE MULTI-KEY GATEWAY RUNNING ON http://127.0.0.1:8080/v1
[12:00:01] 📥 [REQUEST] POST /v1/chat/completions | Model: qwen3.8-max | 🔑 Key #1: sk-xkiro-acc1...
[12:00:02] ⚠️ [RATE-LIMIT] HTTP 429 (Too Many Requests) received from upstream router.
[12:00:02] 🔄 [FAILOVER] Key #1 on 60s cooldown ➔ ⚡ Switching to Key #2 in 0.38ms!
[12:00:02] 📥 [REQUEST] POST /v1/chat/completions | Model: qwen3.8-max | 🔑 Key #2: sk-xkiro-acc2...
[12:00:03] 🚀 [STREAMING] 200 OK ➔ Raw SSE stream delivering code to Trae with ZERO interruptions!| Feature & Metric | Standard Trae | Cursor Pro ($20/mo) | Windsurf ($20/mo) | ⚡ Trae-GodMode ($0) |
|---|---|---|---|---|
| Monthly Subscription | Free Tier Limits | $20 / month | $20 / month | 🟢 $0 (100% Free & Unlimited) |
| Max Working Context | 200,000 Tokens | 128k – 200k | 128k – 200k | 🟢 1,048,576 Tokens (1M) |
| Reasoning Control | Hidden (null) |
Static Toggle | Static Toggle | 🟢 6 Levels (none → max) |
| Terminal Autonomy | Requires Confirmation | Partial Permission | Partial Permission | 🟢 100% Zero-Prompt Auto-Run |
| MCP Tool Execution | Confirmation Modal | Prompts User | Prompts User | 🟢 100% Auto-Approved |
| Model Customizer | Disabled Fields | Fixed | Fixed | 🟢 100% Unlocked Studio |
| Gateway Metric | One-API (20k+ ⭐) | LiteLLM (16k+ ⭐) | OmniRouter | ⚡ Trae-GodMode Gateway |
|---|---|---|---|---|
| Failover Switching Latency | ~8 – 15 ms | ~15 – 30 ms | ~5 – 10 ms | 🟢 < 0.5 ms (Sub-millisecond) |
| Memory Footprint (RAM) | ~280 MB | ~200 MB | ~120 MB | 🟢 < 12 MB (Ultra-Lightweight) |
| SSE Streaming Handling | Buffered / Chunked | Re-encoded | Buffered | 🟢 Raw Byte Pass-Through |
| Tool-Calling Headers | Partial | Can Drop Headers | Partial | 🟢 100% Native Byte Integrity |
| Setup & Dependencies | Docker + MySQL + Redis | 50+ Python Packages | Node + SQLite | 🟢 Zero Dependencies (Python StdLib) |
| Launch Speed | 30–60 minutes | 15–30 minutes | 10–15 minutes | 🟢 Instant (< 2 seconds via .BAT) |
| Your Real-World Scenario | Best Solution | Why |
|---|---|---|
| Enterprise with 100+ Team Members | One-API / LiteLLM | Provides multi-tenant RBAC, billing dashboards, and team token quotas. |
| Solo Developer or Pair-Programmer | ⚡ Trae-GodMode | Pure zero-latency speed, no Docker/DB setup, runs locally with 12MB RAM. |
| Total IDE Superpowers (1M Context + Autonomy) | ⚡ Trae-GodMode | The only open-source tool that patches the IDE runtime itself. |
Trae-GodMode supports any endpoint compatible with the OpenAI /v1/chat/completions API specification:
| Category | Provider / Router | Base URL | Top Supported Models | References |
|---|---|---|---|---|
| ⚡ Free Routers | xKiro AI | https://api.xkiro.com/v1 |
qwen/qwen3.8-max:free, deepseek-v3, deepseek-r1 |
Qwen LM Specs |
| Bynara | https://router.bynara.id/v1 |
minimax-m3-free, qwen3.8-27b, glm-5.3-flash |
GLM Architecture | |
| KiraAI | https://kiraai.vn/v1 |
mimo-v2.5, tencent-hy3-free, kimi-k2.5 |
Moonshot Kimi Docs | |
| 💎 Premium APIs | OpenRouter | https://openrouter.ai/api/v1 |
anthropic/claude-4, google/gemini-3.1-pro |
OpenRouter API Docs |
| DeepSeek Official | https://api.deepseek.com/v1 |
deepseek-chat, deepseek-reasoner |
DeepSeek Technical Specs | |
| Groq Cloud | https://api.groq.com/openai/v1 |
llama-3.3-70b-versatile, mixtral-8x22b |
Groq LPU Acceleration | |
| OpenAI | https://api.openai.com/v1 |
gpt-4o, gpt-4.1, o3-mini |
OpenAI Reference | |
| 💻 Local LLMs | Ollama / vLLM / LMStudio | http://127.0.0.1:11434/v1 |
qwen2.5-coder:32b, deepseek-coder:33b |
Ollama Docs • vLLM Project |
git clone https://github.com/Shivam990q/Trae-GodMode.git
cd Trae-GodModeDouble-click 1_CLICK_APPLY_PATCHES.bat
🛡️ Auto-detects your Trae installation across any Windows drive (
C:,D:,E:), backs up your original files, applies patches, and restarts Trae instantly.
Create your local keys.json (or copy from config/keys.example.json):
{
"routes": {
"xkiro": {
"upstream": "https://api.xkiro.com/v1",
"models": ["qwen/qwen3.8-max:free", "*"],
"keys": [
"sk-xkiro-account1-key...",
"sk-xkiro-account2-key...",
"sk-xkiro-account3-key..."
]
}
}
}Double-click START_KEY_POOLER.bat, set Trae's Custom URL to http://127.0.0.1:8080/xkiro/v1, and start coding unlimitedly!
Trae-GodMode/
├── 📂 .github/ # CI/CD Workflows & Interactive Issue Forms
│ ├── workflows/ci.yml # Automated cross-device testing on every push
│ └── ISSUE_TEMPLATE/ # Bug report & feature request templates
│
├── 📂 core/ # Core Modular Python Engine
│ ├── patcher.py # AST & regex bundle modification engine
│ ├── gateway.py # Zero-latency multi-key proxy & failover engine
│ └── models.py # Capability matrix for uncapped models
│
├── 📂 config/ # Safe Configuration Templates
│ ├── keys.example.json # Multi-account pool routing template
│ └── permissions.template.json # 100% autonomous global permission profile
│
├── 📂 docs/ # In-Depth Technical Guides
│ ├── ARCHITECTURE.md # Deep-dive system specifications & AST hooks
│ ├── MULTI_ACCOUNT_GUIDE.md # How to pool 50+ free keys for infinite coding
│ ├── BENCHMARKS.md # Latency & throughput profiling data
│ ├── TROUBLESHOOTING.md # Common questions, FAQs & fixes
│ ├── README_DESIGN_GUIDE.md # Masterclass formula for top 0.01% GitHub READMEs
│ └── ZH_CN_GUIDE.md # Chinese language guide (中文体验指南)
│
├── 📂 scripts/ # Standalone CLI Scripts
│ ├── apply_all_patches.py # Universal dynamic patcher script
│ ├── key_pooler.py # Standalone gateway runner
│ └── verify_installation.py # Cross-device self-test verification suite
│
├── ⚡ 1_CLICK_APPLY_PATCHES.bat # 1-Click Patcher Launcher (Root entry point)
├── 🚀 START_KEY_POOLER.bat # 1-Click Gateway Launcher (Root entry point)
├── 📜 CHANGELOG.md # Detailed release notes
├── 🤝 CONTRIBUTING.md # Contribution guidelines
├── 🔒 SECURITY.md # Local-only privacy & key protection policy
├── ⚖️ LICENSE # Official MIT License
└── 📖 README.md # You are here
📖 Click to explore full documentation library
- 🏛️ System Architecture & Subsystems — Learn how AST patches intercept bundle hooks in memory.
- 🔑 Multi-Account Key Pooling Guide — Step-by-step setup to bypass rate-limits across all free routers.
- 📊 Latency & Throughput Benchmarks — In-depth profiling comparing GodMode with LiteLLM and One-API.
- 🛠️ Troubleshooting & FAQ — Answers to updates, rollbacks, and connection setup.
- 🎨 Master README Design Playbook — The exact formula and templates to build god-tier READMEs.
- 🇨🇳 中文使用指南 (Chinese Guide) — 简体中文快速上手与多账号配置教程。
Contributions, feature suggestions, and pull requests are welcomed!
See CONTRIBUTING.md to get started.
# Fork & clone
git checkout -b feature/AmazingSuperpower
python scripts/verify_installation.py
git commit -m "feat: add AmazingSuperpower"
git push origin feature/AmazingSuperpowerDistributed under the MIT License. See LICENSE for details.
Built with ❤️ for the global AI coding community.
If Trae-GodMode leveled up your development workflow, please give it a ⭐ on GitHub!