Recall Alpha Agent

Forget the noise. The real action happens before the hype.

🚀 Overview

The Recall Alpha Agent is an advanced AI-driven system that combines crypto trading intelligence with persistent memory storage. It operates in two primary capacities:

1. Alpha-Gathering Twitter Agent:

   • Monitors selected Twitter profiles for new follows.
   • Extracts potential token mints from new follows' bios.
   • Queries Raydium API for token liquidity pool data.
   • Stores collected data in PostgreSQL/SQLite databases.
   • Detects early memecoin opportunities and generates alpha tweets.

2. Recall Plugin for AI Memory:

   • Integrates with Recall storage to persist chain-of-thought (CoT) logs.
   • Syncs reasoning history to Recall buckets.
   • Fetches historical thought logs to enhance inference.
   • Manages persistent memory through Recall API.

🛠 Alpha Service Workflow

The Alpha Service runs as an Express server in a continuous loop, monitoring specific Twitter accounts for potential trading signals.

🛠 Recall Service Workflow

The Recall Service runs as an Express server in a continuous loop, periodically performing batch syncing jobs, transferring chain-of-thought logs from the local database to Recall.

🔄 Flow of Operations

1. The agent scrapes selected Twitter profiles for new follows.
2. It checks the bio of new follows for token mints.
3. If a mint is found, it queries Raydium API for liquidity pool data.
4. The collected alpha data is logged into the alpha_analysis table.
5. A separate Twitter agent queries alpha_analysis and tweets about new opportunities.
6. The system continuously refines scanning intervals based on activity levels.

🛠 Key Components

Component	Function
Twitter Scraper	Extracts new follows from selected accounts.
Raydium API	Fetches liquidity and pool data for detected tokens.
Database Logging	Stores alpha signals in a structured database.
Automated Tweeting	Posts insightful tweets based on analysis.
Recall Integration	Stores chain-of-thought logs for AI reasoning.

🔥 Modified Twitter Client

The Twitter posting client has been updated to:

• Periodically check the alpha_analysis table.
• Detect new token opportunities.
• Generate engaging tweets based on collected data.
• Maintain a structured posting schedule.

🔥 Modified Database Adapters

The SqliteDatabaseAdapter and PostgresDatabaseAdapter have been updated to accommodate Alpha Agent's needs, including:

• Modified logs table and corresponding queries to keep track of synced logs to Recall
• New twitter_following and related queries to keep track of profiles the users Alpha Agent is observing are following, and which ones are new
• New alpha_analysis table to track collected Alpha for our agent to reference when creating Twitter posts

📌 Recall Plugin Functionality

This agent also provides Recall storage as a plugin for Eliza AI agents, allowing them to:

• ✅ Store and retrieve chain-of-thought logs.
• ✅ Persist agent context across sessions.
• ✅ Sync long-term memory with Recall buckets.
• ✅ Purchase additional storage credits.

🔄 Recall Storage Flow

1. Agent generates reasoning logs and writes them locally.
2. Logs are periodically uploaded to Recall storage.
3. Before inference, past logs are retrieved for context.
4. The agent uses historical thought chains to improve decision-making.

📌 Key Features & Actions

Action	Trigger Format	Description
Create Bucket	"Create a bucket named 'logs'"	Creates or retrieves a Recall bucket.
List Buckets	"Show my Recall buckets"	Fetches available storage buckets.
Store Object	"Add object 'file.txt' to bucket 'backup'"	Saves an object (file, data) in a bucket.
Retrieve Data	"Get object 'data.json' from bucket 'logs'"	Downloads stored objects from Recall.
Buy Credits	"Buy 2 Recall credits"	Purchases more Recall storage credits.

📌 Environment Variables (.env)

Configure these settings to enable full functionality:

# Required APIs and Credentials
OPENAI_API_KEY=your-api-key
SOL_PRIVATE_KEY=your-solana-key

# Twitter Bot Credentials
TWITTER_USERNAME=your-twitter-username
TWITTER_PASSWORD=your-twitter-password
TWITTER_EMAIL=your-twitter-email

# Recall Storage
RECALL_PRIVATE_KEY="your-private-key"
RECALL_BUCKET_ALIAS="your-default-bucket"
RECALL_COT_LOG_PREFIX="cot/"

# Database Configuration (Postgres/SQLite)
POSTGRES_URL=your-postgres-url

# Twitter Monitoring Targets
TWITTER_ACCOUNTS=comma-separated-list

# Posting Schedule
TWITTER_DRY_RUN=false
POST_INTERVAL_MIN=10
POST_INTERVAL_MAX=30

# AI Configuration
USE_OPENAI_EMBEDDING=TRUE
SERVER_PORT=3000
DAEMON_PROCESS=false
TWITTER_TARGET_USERS=

🚀 Running the Services

1️⃣ Setup & Install

To ensure smooth operations and reduce the possibility of dependency errors, please ensure you're using the following node and pnpm versions:

pnpm -v 9.15.4
node -v v22.11.0

Next, install:

pnpm i

To start up only the main agent loop (and not the Alpha-gathering service or Recall), run:

pnpm start

2️⃣ Start the Alpha Monitoring Service

pnpm run start:alpha

3️⃣ Start the Recall Service

pnpm run start:recall

Start all three services in the same terminal

Use the following if you prefer running all three services in the same terminal, instead of running in separate terminals using the three steps above.

pnpm run start:all

🏆 Why Use This Agent?

• 🚀 Early Crypto Alpha – Detect promising tokens before they trend.
• 🤖 Smart Tweeting – Automatically analyze & tweet insights.
• 🧠 Persistent AI Memory – Integrates Recall storage for long-term reasoning.
• 💡 Automated Data Processing – Fully autonomous trading intelligence.

This agent ensures you stay ahead of the market by combining Twitter intelligence, liquidity monitoring, and AI-powered insights into a single automated workflow. 🎯