Fenix AI 交易机器人 是 AI Skill Hub 本期精选Agent工作流之一。综合评分 8.0 分,整体质量较高。我们强烈推荐将其纳入你的 AI 工具库,帮助提升工作效率。
Fenix AI 交易机器人 是一套完整的 AI Agent 自动化工作流方案。通过可视化的节点编排,将复杂的多步骤任务拆解为清晰的自动化流程,实现全程无人值守的智能处理。支持与数百种外部服务和 API 无缝集成,适合构建数据处理管线、业务自动化和 AI 辅助决策系统。
Fenix AI 交易机器人 是一套完整的 AI Agent 自动化工作流方案。通过可视化的节点编排,将复杂的多步骤任务拆解为清晰的自动化流程,实现全程无人值守的智能处理。支持与数百种外部服务和 API 无缝集成,适合构建数据处理管线、业务自动化和 AI 辅助决策系统。
# 方式一:pip 安装(推荐)
pip install fenixai_tradingbot
# 方式二:虚拟环境安装(推荐生产环境)
python -m venv .venv
source .venv/bin/activate # Windows: .venv\Scripts\activate
pip install fenixai_tradingbot
# 方式三:从源码安装(获取最新功能)
git clone https://github.com/Ganador1/FenixAI_tradingBot
cd FenixAI_tradingBot
pip install -e .
# 验证安装
python -c "import fenixai_tradingbot; print('安装成功')"
# 命令行使用
fenixai_tradingbot --help
# 基本用法
fenixai_tradingbot input_file -o output_file
# Python 代码中调用
import fenixai_tradingbot
# 示例
result = fenixai_tradingbot.process("input")
print(result)
# fenixai_tradingbot 配置文件示例(config.yml) app: name: "fenixai_tradingbot" debug: false log_level: "INFO" # 运行时指定配置文件 fenixai_tradingbot --config config.yml # 或通过环境变量配置 export FENIXAI_TRADINGBOT_API_KEY="your-key" export FENIXAI_TRADINGBOT_OUTPUT_DIR="./output"
Data-integrity release — v2.6 is built around one flagship discovery and fix: live indicators were being computed on partial-candle snapshots instead of closed candles. Everything else hardens the stack around it: smarter NanoFenix self-monitoring, a rebuilt vision pipeline, prompt-injection defenses, and a dashboard that finally shows why the engine did (or didn't) trade.
Reliability-focused release — v2.5 brings short-timeframe latency work, a complete performance optimisation pass, NanoFenix v3.5 as a first-class companion signal, DeepSeek v4 cloud experiments, and a full suite of live/paper reliability fixes.
Complete architectural overhaul - Migrated from CrewAI to LangGraph for more robust and flexible agent orchestration.
| Feature | v1.0 (June 2025) | v2.0 (December 2025) |
|---|---|---|
| **Orchestration** | CrewAI | LangGraph (State Machine) |
| **Memory System** | Basic TradeMemory | [ReasoningBank](https://arxiv.org/abs/2509.25140) + LLM-as-Judge |
| **Visual Analysis** | Static screenshots | Chart Generator + Playwright TradingView Capture |
| **LLM Providers** | Ollama only | Ollama, MLX, Groq, HuggingFace |
| **Frontend** | Flask Dashboard | React + Vite + TypeScript |
| **Agent Weighting** | Static | Dynamic (performance-based) |
| **Security** | Basic | SecureSecretsManager + Path Validation |
| **Real-time** | Polling | WebSocket + Socket.IO |
127.0.0.1 by default, demo users are gated, and secrets scanning is part of the developer workflow.RELEASE_CHECKLIST.md before publishing. Dev-focused run instructions are in DEVELOPMENT.md.docs/archives/reports/.docs/security/docs/security/DEMO_CREDENTIALS.md.| Requirement | Version | Notes |
|---|---|---|
| Python | 3.10+ | 3.11 recommended |
| Node.js | 18+ | For frontend |
| Ollama | Latest | Local LLM inference |
| RAM | 16GB+ | 32GB for larger models |
| GPU | Optional | CUDA for faster inference |
| Apple Silicon | M1/M2/M3 | MLX support for optimized inference |
pip install -e ".[dev,vision,monitoring]"
ollama pull qwen3:8b ```
pip install -e ".[dev]"
```bash
```bash cp .env.example .env
```bash
pre-commit install
context = reasoning_bank.get_relevant_context( agent_name="technical_analyst", current_prompt=market_analysis_prompt, limit=3 )
python -m venv .venv source .venv/bin/activate # Linux/Mac
cp .env.example .env
python run_fenix.py --help
python run_fenix.py # Paper trading (default)
python run_fenix.py --symbol ETHUSDT # Different symbol
python run_fenix.py --timeframe 5m # Different timeframe
python run_fenix.py --no-visual # Disable visual agent
python run_fenix.py --mode live --allow-live # Live trading (⚠️ real money)
---
trading:
symbol: BTCUSDT
timeframe: 15m
max_risk_per_trade: 0.02
agents:
enable_technical: true
enable_qabba: true
enable_visual: true # Requires vision model
enable_sentiment: true # Requires news APIs
technical_weight: 0.30
qabba_weight: 0.30
consensus_threshold: 0.65
active_profile: "all_local" # Options: all_local, mixed_providers, mlx_optimized, all_cloud
all_local:
technical:
provider_type: "ollama_local"
model_name: "qwen3:8b"
temperature: 0.1
| Variable | Description | Default |
|---|---|---|
BINANCE_API_KEY | Binance API key | - |
BINANCE_SECRET_KEY | Binance secret key | - |
LLM_PROFILE | LLM provider profile to use | all_local |
GROQ_API_KEY | Groq API key (for cloud inference) | - |
HF_TOKEN | HuggingFace token | - |
ALLOW_EXPOSE_API | Allow API to bind to all interfaces | false |
CREATE_DEMO_USERS | Enable demo user creation | false |
LLM_ALLOW_NOOP_STUB | Fallback to noop LLM for testing | 0 |
ENABLE_VISUAL_AGENT | Enable chart analysis agent | true |
ENABLE_SENTIMENT_AGENT | Enable news/social analysis | true |
---
python run_fenix.py --api
docker compose up -d --build
docker compose --profile monitoring up -d --build ```
Docker defaults to Python 3.12, publishes the API only on 127.0.0.1:8001, and keeps Redis internal to the Compose network.
---
| Fix | Details |
|---|---|
| **Prompt formatting bug** | The visual prompt contained a literal JSON example with unescaped braces — str.format() crashed before the image ever reached the model. |
| **Chart generator accuracy** | The chart labeled an SMA 50 as "EMA 50", had no pivot overlay, and drew a false vertical line from SuperTrend warmup zeros. Now: true EMA 50, R3..S3 pivot levels, NaN warmup. |
| **Chart metadata in the prompt** | The vision model receives the real candle count plus numeric EMA/Bollinger/VWAP/SuperTrend/pivot values alongside the image. |
| **Compact JSON contract** | Single-line, ~300-char response schema with a raised token budget — eliminates truncated-JSON retry loops. |
| Model | Role | Notes |
|---|---|---|
**DeepSeek v4 Flash** (deepseek-v4-flash:cloud) | Technical / Decision | Fast, cost-efficient cloud inference |
**DeepSeek v4 Pro** (deepseek-v4-pro:cloud) | Full pipeline | Highest-accuracy cloud option tested |
| **cogito-2.1:671b-cloud** | QABBA | Benchmark winner: 75–80% directional accuracy |
| **nemotron-3-nano:30b-cloud** | Technical + Decision | 66.7% accuracy; most active decision model |
| **glm-5:cloud** | Risk Manager | 77.8% activity rate, score 0.504 in benchmark |
127.0.0.1 by default to avoid accidental public exposure. To bind to all interfaces intentionally, set ALLOW_EXPOSE_API=true.CREATE_DEMO_USERS=true for local development.DEFAULT_DEMO_PASSWORD and DEFAULT_ADMIN_PASSWORD may be used for local testing; avoid using them in production.DEVELOPMENT.md and RELEASE_CHECKLIST.md to help developers follow the release process and avoid secrets leaks.docs/archives/reports/ to reduce root clutter.---
┌─────────────────────────────────────────────────────────────────────────────┐
│ FENIX AI v2.5 RC │
├─────────────────────────────────────────────────────────────────────────────┤
│ │
│ ┌─────────────┐ ┌──────────────────────────────────────────────────┐ │
│ │ Frontend │◄──►│ FastAPI + Socket.IO │ │
│ │ React/Vite │ │ (Real-time) │ │
│ └─────────────┘ └────────────────────┬─────────────────────────────┘ │
│ │ │
│ ┌───────────────────────────────────────▼──────────────────────────────┐ │
│ │ TRADING ENGINE │ │
│ │ ┌─────────────────────────────────────────────────────────────────┐ │ │
│ │ │ LangGraph Orchestrator │ │ │
│ │ │ (State Machine) │ │ │
│ │ └─────────────────────────────────────────────────────────────────┘ │ │
│ │ │ │ │ │ │ │
│ │ ┌─────▼─────┐ ┌─────▼─────┐ ┌─────▼─────┐ ┌─────▼─────┐ │ │
│ │ │ Technical │ │ Visual │ │ Sentiment │ │ QABBA │ │ │
│ │ │ Agent │ │ Agent │ │ Agent │ │ Agent │ │ │
│ │ └─────┬─────┘ └─────┬─────┘ └─────┬─────┘ └─────┬─────┘ │ │
│ │ │ │ │ │ │ │
│ │ ┌─────▼──────────────▼──────────────▼──────────────▼─────┐ │ │
│ │ │ Decision Agent + Risk Manager │ │ │
│ │ │ (Dynamic Weighting + LLM-as-Judge) │ │ │
│ │ └────────────────────────┬───────────────────────────────┘ │ │
│ └──────────────────────────────┼───────────────────────────────────────┘ │
│ │ │
│ ┌──────────────────────────────▼───────────────────────────────────────┐ │
│ │ MEMORY LAYER │ │
│ │ ┌─────────────────┐ ┌─────────────────┐ ┌─────────────────────┐ │ │
│ │ │ ReasoningBank │ │ Trade Memory │ │ LLM-as-Judge │ │ │
│ │ │ (Semantic Search)│ │ (History) │ │ (Self-Evaluation) │ │ │
│ │ └─────────────────┘ └─────────────────┘ └─────────────────────┘ │ │
│ └──────────────────────────────────────────────────────────────────────┘ │
│ │
│ ┌──────────────────────────────────────────────────────────────────────┐ │
│ │ EXECUTION LAYER │ │
│ │ ┌─────────────────┐ ┌─────────────────┐ ┌─────────────────────┐ │ │
│ │ │ Binance Client │ │ Order Executor │ │ Market Data │ │ │
│ │ │ (REST + WS) │ │ (Paper/Live) │ │ (Real-time) │ │ │
│ │ └─────────────────┘ └─────────────────┘ └─────────────────────┘ │ │
│ └──────────────────────────────────────────────────────────────────────┘ │
│ │
└─────────────────────────────────────────────────────────────────────────────┘
pytest tests/test_integration.py -v
功能齐全的AI交易机器人,支持多种服务提供者
AI Skill Hub 为第三方内容聚合平台,本页面信息基于公开数据整理,不对工具功能和质量作任何法律背书。
建议在沙箱或测试环境中充分验证后,再部署至生产环境,并做好必要的安全评估。
✅ Apache 2.0 — 宽松开源协议,可商用,需保留版权声明和 NOTICE 文件,含专利授权条款。
经综合评估,Fenix AI 交易机器人 在Agent工作流赛道中表现稳健,质量优秀。如果你已有明确的使用需求,可以直接上手体验;如果还在评估阶段,建议对比同类工具后再做决策。
| 原始名称 | FenixAI_tradingBot |
| Topics | aiautonomous-agentspythontrading-bot |
| GitHub | https://github.com/Ganador1/FenixAI_tradingBot |
| License | Apache-2.0 |
| 语言 | Python |
收录时间:2026-07-03 · 更新时间:2026-07-03 · License:Apache-2.0 · AI Skill Hub 不对第三方内容的准确性作法律背书。
选择 Agent 类型,复制安装指令后粘贴到对应客户端