Discord Crypto Analyzer

A powerful Discord bot for real-time cryptocurrency market analysis using advanced AI models.

Try It Live

Join our Discord server to see the bot in action and test features in real time:

👉 Discord Invitation Link

Channel: #ai-analysis

Overview

Discord Crypto Analyzer provides real-time cryptocurrency analysis directly to your Discord server. Using advanced AI models, it generates detailed reports, forecasts, and detects key market patterns—perfect for traders and enthusiasts seeking instant, data-driven insights.

Features

• Real-time Market Analysis: Analyze any cryptocurrency trading pair with a simple command
• AI-Powered Insights: Utilizes advanced AI models to interpret complex market data
• Visual Chart Analysis: Automatically generates candlestick charts that AI models analyze for enhanced pattern recognition and market insights
• Technical Indicators: Comprehensive analysis using RSI, MACD, Bollinger Bands, Supertrend, and many more.
• Multi-Timeframe Analysis: Examines data across multiple timeframes (1h, 2h, 4h, 6h, 8h, 12h, 1d).
• Flexible Timeframe Selection: Analyze markets on your preferred timeframe via Discord commands or config
• Support & Resistance Levels: Identifies key price levels.
• Price Targets: Provides short-term and medium-term price targets
• Sentiment Analysis: Incorporates Fear & Greed index for market sentiment
• Detailed Reports: Generates in-depth HTML reports for comprehensive analysis
• Anti-Spam Protection: Built-in measures to prevent channel spam and abuse
• Context-Aware Analysis: Uses RAG (Retrieval-Augmented Generation) for enhanced market context
• Multi-Language Support: Analyze markets in multiple languages
• Auto-Categorization: Automatically identifies and categorizes crypto assets
• Multi-Exchange Support: Fetches data from multiple exchanges including Binance, KuCoin, Gate.io, MEXC, and Hyperliquid
• Divergence Detection: Identifies bullish and bearish divergences across multiple indicators
• Streaming Analysis: Real-time streaming responses when using local AI models for faster feedback
• Automatic Cleanup: Messages and files automatically expire and are cleaned up after a configurable period
• Fault-Tolerant Design: Multiple AI model fallbacks for continuous operation
• Smart Caching: Optimized data retrieval with minimal API calls
• Educational Resources: Automated linking to indicator explanations and definitions
• Pattern Recognition: Detects common chart patterns like head and shoulders, wedges, etc.
• Market Metrics: Calculates key metrics across various timeframes
• Weekly Macro Trends: Analyzes 200-week SMA for institutional-grade macro trend analysis with cycle phase identification

Installation

Prerequisites

• Python 3.13 or higher
• Discord Bot Token
• OpenRouter API Key or Local LM Studio setup
• CryptoCompare API Key (for market data)
• Google Studio API Key (optional, for Gemini models via official Google GenAI SDK)

Steps

These installation instructions assume you have a local copy of the repository (clone or download). If you're using a fork or mirror, follow the same steps below to create a virtual environment and install dependencies.

1. Clone or download the repository to your local machine (or use your fork).

2. (Recommended) Create a Python virtual environment:

python -m venv .venv

Activate the virtual environment:

• On Windows (PowerShell):

  .venv\Scripts\Activate.ps1

• On Windows (cmd):

  .venv\Scripts\activate.bat

• On macOS/Linux:

  source .venv/bin/activate

3. Install the required dependencies:

pip install -r requirements.txt

4. Create configuration files:

   • Copy keys.env.example to keys.env
   • Set your API keys and Discord settings in the keys.env file
   • Adjust other settings in config/config.ini as needed

5. Run the bot:

python start.py

Configuration

The configuration is split across two files:

keys.env

Contains sensitive information that should not be committed to version control:

• BOT_TOKEN_DISCORD: Your Discord bot token
• OPENROUTER_API_KEY: API key for OpenRouter AI services
• GOOGLE_STUDIO_API_KEY: API key for Google Studio (Gemini) models
• CRYPTOCOMPARE_API_KEY: API key for CryptoCompare data
• GUILD_ID_DISCORD: ID of your Discord server
• MAIN_CHANNEL_ID: Main channel for bot communication
• TEMPORARY_CHANNEL_ID_DISCORD: Channel for temporary file uploads

config/config.ini

Contains non-sensitive configuration that can be safely committed.

Timeframe Configuration

The bot supports flexible timeframe analysis from 1 hour to 1 day:

Supported Timeframes: 1h, 2h, 4h, 6h, 8h, 12h, 1d

Setting Default Timeframe

In config/config.ini:

[general]
timeframe = 1h  # Change to 4h, 1d, etc.

This sets the default timeframe used when users don't specify one in their command.

Per-Analysis Timeframe Override

Users can specify timeframe directly in Discord commands:

!analyze BTC/USDT          → Uses config default (1h)
!analyze BTC/USDT 4h       → Analyzes on 4-hour timeframe
!analyze BTC/USDT 1d       → Analyzes on daily timeframe
!analyze BTC/USDT 4h Polish → 4-hour timeframe in Polish

How It Works:

• Dynamic Calculations: All technical indicators, period calculations, and candle progress tracking automatically adapt to the selected timeframe
• Smart Data Fetching: System automatically adjusts candle limits to maintain ~30 days of historical data across all timeframes
• Real-time Flexibility: Users don't need to wait for bot restart to try different timeframes

Choosing Your Timeframe:

• 1h-2h: Short-term trading, high granularity, more frequent signals
• 4h-6h: Swing trading, balanced view, medium-term trends
• 8h-12h: Longer swing positions, reduced noise
• 1d: Position trading, macro trends, most stable signals

How AI Provider Selection Works

Use the provider setting in the [ai_providers] section of config/config.ini to select the AI provider. This section explains the available options and exact runtime behavior.

• provider (string) — valid values and behavior:

  • "local": Use LM Studio only (local models). LM Studio is the only provider used; streaming is supported when LM Studio is available. If LM Studio fails, the request will fail — there is no automatic fallback.
  • "googleai": Use Google AI Studio only (Gemini models). Google is the single provider; no automatic fallbacks to other providers on failure.
  • "openrouter": Use OpenRouter only. OpenRouter is the single provider; no automatic fallbacks to other providers on failure.
  • "all": Use the fallback chain. In this mode the runtime will try providers in this order until one returns a valid response:
    1. Google AI Studio
    2. LM Studio (local)
    3. OpenRouter

  When provider is set to "all", the system preserves the original fallback behavior. For single-provider settings ("local", "googleai", "openrouter") there is no cross-provider fallback.

• lm_studio_base_url: URL for LM Studio (default: "http://localhost:1234/v1")

• lm_studio_model: Model identifier for LM Studio (default: "local-model")

• OPENROUTER_BASE_URL: Base URL for OpenRouter API (default: "https://openrouter.ai/api/v1")

• OPENROUTER_BASE_MODEL: Default OpenRouter model identifier (e.g. "google/gemini-2.5-flash")

• OPENROUTER_FALLBACK_MODEL: Fallback model identifier for OpenRouter

• GOOGLE_STUDIO_MODEL: Default Google model identifier (e.g. "gemini-2.5-flash")

• TIMEFRAME: Default timeframe for analysis (default: "1h")

• LOG_DIR: Directory for log files (default: "logs")

• LOGGER_DEBUG: Enable debug logging

• ANALYSIS_COOLDOWN_COIN: Time between analyses of the same coin (in seconds)

• ANALYSIS_COOLDOWN_USER: Time between user requests (in seconds)

• RAG_UPDATE_INTERVAL_HOURS: How often to update the RAG (Retrieval-Augmented Generation) system

• SUPPORTED_LANGUAGES: Dictionary of supported languages for analysis output

Usage

Commands

• !analyze <symbol> [timeframe] [language]: Analyze a trading pair with optional timeframe and language
  • Basic: !analyze BTC/USDT (uses config default timeframe, English)
  • With timeframe: !analyze BTC/USDT 4h (4-hour timeframe, English)
  • With language: !analyze BTC/USDT Polish (config default timeframe, Polish)
  • Full: !analyze BTC/USDT 1d Polish (daily timeframe, Polish)
• !cleanup: Force cleanup of expired messages (owner only)
• !unmute <user>: Unmute a user that was muted for spamming

Supported Timeframes: 1h, 2h, 4h, 6h, 8h, 12h, 1d

Admin: Custom AI Provider & Model Selection

Administrators can override the default AI provider and model for analysis. This is useful for testing different models or using specific providers:

!analyze <symbol> [timeframe] [language] <provider> <model>

Available Providers:

• googleai — Google AI Studio (Gemini models)
• openrouter — OpenRouter (access to 500+ models: GPT, Claude, Gemini, DeepSeek, etc.)
• local — LM Studio (local models)
• all — Fallback chain (tries: Google AI → LM Studio → OpenRouter)

Admin Command Examples:

# Use specific Gemini model
!analyze BTC/USDT googleai gemini-2.5-pro

# Use OpenRouter with GPT model
!analyze BTC/USDT openrouter openai/gpt-4o

# Use custom local model
!analyze ETH/USDT 4h local llama-2-7b

# Full command with timeframe and language
!analyze BTC/USDT 4h openrouter openai/gpt-4-turbo

Admin Benefits:

• No cooldown periods when using provider/model override
• Direct access to all configured AI providers and models
• Useful for testing and comparing different models
• Skip rate limiting restrictions

Future: Tiered Model Access for Paid Users

As the platform grows with a paid userbase, the bot will support tiered model access:

• Free Tier: Default model (e.g., gemini-flash-latest)

  • Standard analysis quality
  • Shared rate limits
  • Community channels

• Premium Tier: Access to advanced models (e.g., GPT-5, Claude 4.5 Sonnet (recommended), Gemini 2.5 Pro)

  • Superior analysis quality and reasoning
  • Higher accuracy for pattern recognition
  • Dedicated rate limits
  • Priority processing
  • Advanced features

This tiered approach will allow casual users to enjoy the bot's capabilities while providing paid users with access to state-of-the-art AI models for more accurate financial analysis. Implementation details and pricing will be determined as the community grows.

Note: There are no !shutdown or !restart Discord commands. To stop the running bot use the console keyboard shortcut q (Windows console) or send an OS signal (SIGINT/SIGTERM) to the process.

Example Commands

# Basic analysis (default timeframe from config)
!analyze BTC/USDT

# Different timeframes
!analyze BTC/USDT 4h
!analyze BTC/USDT 1d  
!analyze ETH/USDT 12h

# With language
!analyze BTC/USDT Polish
!analyze BTC/USDT 4h Spanish
!analyze BTC/USDT 1d French

# More combinations
!analyze BTC/USDT German
!analyze ETH/USDT 4h Chinese
!analyze SOL/USDT 1d Japanese
!analyze BNB/USDT 12h Russian

Supported languages: English, Polish, Spanish, French, German, Chinese, Japanese, Russian

The bot will process the request and provide:

1. A confirmation message that analysis is in progress
2. A detailed embed with the analysis results including trend direction, strength.
3. An HTML file with the complete in-depth analysis

Architecture

The bot is structured into a well-organized architecture with several key components:

• Core Application: Handles initialization, shutdown, and coordination between components
• Market Analysis Engine: Orchestrates data collection, indicator calculation, and analysis generation
  • Data Collector: Fetches market data from exchanges and APIs
  • Indicator Calculator: Computes technical indicators and detects patterns
  • Result Processor: Interprets data through AI models
  • Analysis Publisher: Distributes results through Discord
• Discord Interface: Manages commands, message handling, and user interactions
  • Command Handler: Processes user commands and coordinates responses
  • Anti-Spam System: Prevents abuse through request rate limiting
  • Reaction Handler: Enables interactive elements through emoji reactions
  • File Handler: Manages temporary file uploads and cleanup
• RAG Engine: Retrieval-Augmented Generation system for enhanced analysis
  • News Database: Maintains indexed repository of recent crypto news
  • Market Context: Provides real-time market data for analysis
  • Periodic Updates: Automatically refreshes data at configured intervals
  • Symbol Recognition: Identifies cryptocurrency symbols in news articles
• Platform APIs: Interfaces with data sources
  • CoinGecko API: Market caps, dominance metrics, and global data
  • CryptoCompare API: News, OHLCV data, and price information
  • Alternative.me API: Fear & Greed index for sentiment analysis
• AI Model Management:
  • Multi-Provider Strategy: Uses local and cloud-based AI models
  • Official Google GenAI SDK: Integrated official Google AI SDK for enhanced reliability
  • Fallback Chain: Gracefully handles API failures and rate limits
  • Response Parsing: Transforms raw AI output to structured data
  • Token Tracking: Monitors and optimizes token usage
• HTML Generator: Creates visual reports of analysis results
• Logger System: Comprehensive logging with error tracking

Technical Analysis System

The bot implements a comprehensive technical analysis system that combines traditional indicators with AI interpretation:

Technical Indicators

• Trend Indicators: ADX, Directional Movement (DI+/DI-), Supertrend
• Momentum Indicators: RSI, MACD, Stochastic Oscillator, Williams %R
• Volume Indicators: MFI, OBV, Chaikin Money Flow, Force Index
• Volatility Indicators: ATR, Bollinger Bands
• Price Indicators: VWAP, TWAP, Simple Moving Averages, Exponential Moving Averages
• Statistical Indicators: Z-Score, Kurtosis, Hurst Exponent

Pattern Recognition

The system automatically detects technical patterns including:

• RSI Patterns: Oversold/overbought conditions, W-bottoms, M-tops
• MACD Patterns: Bullish/bearish crossovers, zero-line crosses
• Divergence Detection: RSI, MACD, and Stochastic divergences (bullish/bearish)
• Crossover Patterns: DI+/DI- crossovers, Supertrend direction changes
• Volatility Patterns: Expansion/contraction, spikes, trend changes
• Support/Resistance Levels: Advanced level calculation with volume confirmation

Additionally, the AI visually analyzes chart images to identify classic price patterns such as head and shoulders, triangles, wedges, channels, and candlestick formations.

Performance Optimization

• Smart Caching: Indicators are cached using data hashing to avoid redundant calculations
• Parallelized Computation: Leverages numpy vectorization for performance
• Progressive Loading: HTML reports load essential components first for better user experience

HTML Report System

The bot generates detailed HTML analysis reports that include:

Interactive Charts

• Candlestick charts with price and volume data, RSI
• Interactive controls for zooming, panning, and data exploration

Content Enrichment

• Smart Indicator Linking: Technical terms automatically link to educational resources
• News Integration: Relevant news articles are referenced and linked
• Source Attribution: News sources are properly cited and accessible
• Responsive Design: Reports display properly on mobile and desktop devices

Analysis Components

• Technical analysis summary
• Trend direction and strength assessment
• Support and resistance levels
• Price target projections
• Timeframe comparisons
• Confidence scores
• Risk assessment metrics

AI Analysis System

The bot leverages advanced AI models to interpret market data and generate insights through a sophisticated prompt engineering system:

Prompt Engineering

• Structured Data Format: Market data is organized into clearly defined sections for optimal AI processing
• Context Enrichment: Current market data is enhanced with historical context and sentiment indicators
• Multi-Timeframe Integration: Analysis spans from hourly to yearly data for comprehensive perspective
• Pattern Highlighting: Notable patterns and setups are automatically emphasized for the AI model
• Guided Analysis Framework: Step-by-step analysis instructions ensure thorough examination of all factors

Response Processing

• JSON Parsing: Extracts structured data from AI responses for standardized display in Discord
• Markdown Formatting: Converts detailed analysis into well-formatted, readable content
• Language Adaptation: Processes analysis in multiple languages while maintaining technical accuracy
• Error Handling: Gracefully handles malformed AI responses with fallback strategies

AI Model Orchestration

• Tiered Provider Strategy: Supports local models (LM Studio), cloud-based providers (Google AI Studio, OpenRouter)
• Configurable Fallback: When set to "all" mode, tries Google AI Studio → LM Studio (local) → OpenRouter
• Streaming Support: Real-time streaming responses available with LM Studio for faster interactive analysis
• Adaptive Selection: Dynamically switches between models based on availability and response quality
• Response Parsing: Transforms raw AI output to structured data
• Token Optimization: Intelligently manages prompt length to balance detail and token usage

RAG System (Retrieval-Augmented Generation)

The bot includes a sophisticated RAG engine that enhances market analysis with relevant context:

• News Indexing: Automatically collects and indexes cryptocurrency news articles
• Market Context: Provides real-time market data including global metrics and sentiment
• Multi-dimensional Search: Finds relevant information across categories, tags, and content
• Content Ranking: Prioritizes information based on recency, relevance, and importance
• Contextual Integration: Seamlessly incorporates retrieved information into AI prompts
• Automated Updates: Periodically refreshes knowledge base with latest market information
• Symbol Recognition: Automatically identifies cryptocurrency mentions in news articles

Example Output

You can find a sample HTML analysis report in the example_html directory:

• example_html/ETHUSDT_analysis_20250915155439.html

This file demonstrates the bot's output for a typical market analysis, including interactive charts and technical indicators.

Example Images

Below are sample images generated by the bot, located in the img directory:

License

MIT

Disclaimer

This bot is for informational and educational purposes only. It does not provide trading signals or financial recommendations, and should not be considered as financial advice. Always do your own research before making investment decisions.