A decentralized AI trading agent platform on Solana, leveraging RIG for LLM interactions and MongoDB for vector storage.

Token Filter Pipeline

The token filter pipeline is a core component that analyzes and filters Solana tokens using the BirdEye API and LLM-based analysis. The pipeline consists of five stages:

1. BirdEye Filter Selection: Intelligent selection of filtering parameters using LLM
2. Token List Retrieval: Fetching and filtering tokens from BirdEye API
3. Market Analysis: Deep analysis of market metrics and trading patterns
4. Metadata Analysis: Evaluation of social signals and development metrics
5. Final Filtering & Storage: Storage of filtered tokens with comprehensive analysis

Features

• Fully automated token analysis and filtering
• Integration with BirdEye API for real-time market data
• MongoDB storage for token recommendations
• LLM-powered market analysis and risk assessment
• Comprehensive scoring system for token evaluation

Technical Stack

• Language: Rust
• APIs: BirdEye API (Solana)
• Database: MongoDB
• LLM Integration: RIG with O1-MINI model
• Data Storage: Vector storage for token analysis

Getting Started

1. Clone the repository
2. Install dependencies:

   cargo build

3. Set up environment variables:

   BIRDEYE_API_KEY=your_api_key
   OPENAI_API_KEY=your_api_key
   MONGODB_URI=your_mongodb_uri

4. Run the token filter example:

   cargo run --example token_filter

Pipeline Configuration

The token filter pipeline can be configured through:

• BirdEye API parameters in src/prompts/token_filter_initial.txt
• MongoDB collection settings in src/services/token_filter.rs
• LLM model selection in environment variables

Data Structures

• BirdeyeFilters: API query parameters
• TokenAnalysis: Token evaluation data
• FilterResponse: Analysis results
• FilterSummary: Market statistics

Next Steps

• Add visualization layer for analysis results
• Implement real-time monitoring
• Enhance social/dev metrics analysis
• Create web dashboard for token recommendations

Contributing

1. Fork the repository
2. Create your feature branch
3. Commit your changes
4. Push to the branch
5. Create a new Pull Request

License

This project is licensed under the MIT License - see the LICENSE file for details.

Overview

Cainam Core is a Rust-based system that implements autonomous AI trading agents, market monitoring, and data analysis for the Solana blockchain. The system features real-time market data processing, automated trading execution, and advanced risk management capabilities.

Key Features

• Real-time market monitoring via Birdeye API
• Blockchain transaction monitoring using Helius webhooks
• Autonomous trading agents with AI-driven decision making
• Advanced risk management and position sizing
• Time-series data storage with TimescaleDB
• Vector similarity search using Qdrant
• Discord and Twitter integration

Prerequisites

• Rust 1.75+ (2021 edition)
• PostgreSQL 15+ with TimescaleDB extension
• Solana CLI tools
• Node.js and npm (for development tools)

Installation

1. Clone the repository:

git clone https://github.com/cainamventures/cainam-core
cd cainam-core

2. Copy the environment template and configure your variables:

cp .env.example .env
# Edit .env with your configuration

3. Install development dependencies:

# Install pre-commit hooks
pre-commit install

# Install required database extensions
psql -c 'CREATE EXTENSION IF NOT EXISTS timescaledb;'

4. Build the project:

cargo build

Configuration

The following environment variables are required:

# Database
DATABASE_URL=postgresql://user:password@localhost/dbname

# Solana
SOLANA_RPC_URL=your_rpc_url
HELIUS_API_KEY=your_helius_key

# APIs
BIRDEYE_API_KEY=your_birdeye_key

# Optional integrations
DISCORD_TOKEN=your_discord_token
TWITTER_API_KEY=your_twitter_key

Project Structure

src/
├── actions/      # External API interactions
├── agent/        # Agent implementations
├── trading/      # Trading logic
├── models/       # Data models
└── services/     # Business logic

Development

Running Tests

# Run all tests
cargo test

# Run specific test suite
cargo test --package cainam-core

Code Style

The project uses rustfmt and clippy for code formatting and linting:

# Format code
cargo fmt

# Run clippy
cargo clippy

Performance Requirements

• Trade execution: < 500ms end-to-end
• Market data updates: < 1s refresh rate
• Signal processing: < 200ms
• Database queries: < 100ms response time

Dependencies

Core dependencies include:

• tokio (async runtime)
• solana-client & solana-sdk (blockchain interaction)
• serde (serialization)
• tokio-postgres (database)
• qdrant-client (vector store)
• rig-core (framework)

Contact

• Author: Matt Gunnin
• Email: matt@cainamventures.com
• Repository: https://github.com/cainamventures/cainam-core

CLI Usage

The cainam CLI tool provides several commands for interacting with the platform:

Get Trending Tokens

cargo run --bin cainam trending

This command displays a list of trending tokens with their current prices, 24h changes, volumes, and market caps.

Get Token Overview

cargo run --bin cainam token <TOKEN_ADDRESS>

Shows detailed information about a specific token, including:

• Current price and market cap
• 24h volume and price change
• Number of holders and active wallets

Analyze Market Signals

cargo run --bin cainam signals <TOKEN_ADDRESS>

Analyzes market signals for a specific token, displaying:

• Signal type (if any)
• Confidence score
• Risk assessment
• Price and volume changes

Monitor Tokens

cargo run --bin cainam monitor <TOKEN_ADDRESS1,TOKEN_ADDRESS2,...> [--interval <SECONDS>]

Continuously monitors specified tokens for market signals. The interval defaults to 300 seconds (5 minutes).

CLI Usage

The token filter pipeline can be run through a user-friendly CLI interface that provides real-time feedback and colored output:

# Run the token filter with continuous monitoring
cargo run --example token_filter_cli filter --continuous

# Run a single token filter pass
cargo run --example token_filter_cli filter

The CLI provides:

• Colored section headers and progress indicators
• Real-time token analysis with detailed metrics
• Visual score bars for different metrics
• Color-coded market signals and risk assessments

Example output includes:

• Token information (name, symbol, price) in bold white
• Prices in green
• Market cap in yellow
• Volume in blue
• Price changes (positive in green, negative in red)
• Score bars (green ≥0.8, yellow ≥0.6, red <0.6)
• Progress spinners for long-running operations

CLI Features

• Progress Tracking: Real-time progress indicators for long-running operations
• Color-Coded Output: Intuitive color scheme for different metrics
• Visual Score Bars: Progress bar style visualization of scores
• Section Headers: Clear organization of output with colored headers
• Market Signals: Detailed market signal information with confidence scores
• Analysis Summary: Overview of token analysis with pass/fail counts

BirdEye V3 API Integration Testing

The feature/birdeye-v3-llm-filter branch implements enhanced token filtering using BirdEye V3 API and LLM-based analysis. To test the implementation:

1. Clone and Setup:

   git clone https://github.com/CainamVentures/cainam-core.git
   cd cainam-core
   git checkout feature/birdeye-v3-llm-filter

2. Environment Setup: Create a .env file with:

   BIRDEYE_API_KEY=<your-key>
   OPENAI_API_KEY=<your-key>
   MONGODB_URI=<your-uri>
   MONGODB_DATABASE=<your-db>
   

3. Build and Run:

   cargo build
   cargo run --example token_filter

4. Verify Operation:

   • Check logs for successful API calls
   • Monitor MongoDB for stored analysis
   • Verify token filtering results

The branch is ready to merge into main. All tests are passing, and the implementation is production-ready.