Setting Up Your Own Twitter Agent with ElizaOS

Deploy a customized AI agent on Twitter

ElizaOS enables the deployment of AI agents capable of automated posting and interaction on Twitter (X). This guide provides the steps to configure and run your own Twitter agent.

Getting Started with ElizaOS

Follow these steps to set up your Twitter agent.

Step 1: Install the ElizaOS CLI

First, install the ElizaOS Command Line Interface (CLI), which provides the necessary tools for interacting with the ElizaOS platform.

Open your terminal and execute the following command:

npm install -g @elizaos/cli@beta

This command installs the ElizaOS CLI globally on your system.

Step 2: Create a New Project

With the CLI installed, create a new project directory:

elizaos create

This command initiates an interactive setup wizard:

• Enter a name for your project (e.g., "my-twitter-agent").
• When prompted for a database, select "pglite" for simplified setup.
• Follow the remaining prompts to configure the project.

Alternatively, create the project non-interactively:

elizaos create my-twitter-agent

Step 3: Configure Environment Variables

Next, configure the agent's Twitter credentials and settings using environment variables.

Open the ElizaOS environment configuration file in a text editor:

# Recommended: Use Cursor (https://cursor.sh)
# Open Cursor, navigate to File > Open, and select the .env file:
# - Windows: C:\Users\YOUR_USERNAME\.eliza\.env
# - Mac/Linux: ~/.eliza/.env

# Alternative editors:
# Windows: notepad C:\Users\YOUR_USERNAME\.eliza\.env
# Mac: open -a TextEdit ~/.eliza/.env
# Linux: gedit ~/.eliza/.env

Add the following lines to the .env file, replacing placeholder values with your actual credentials and desired settings:

# Global environment variables for Eliza
# Use an absolute path, as '~' might not be expanded by all tools
PGLITE_DATA_DIR=/path/to/your/home/.eliza/db

# OpenAI API Key (Optional, if using OpenAI features)
OPENAI_API_KEY=your_openai_api_key_here

# Twitter Configuration
TWITTER_USERNAME=your_twitter_username
TWITTER_PASSWORD=your_twitter_password
TWITTER_EMAIL=your_twitter_email
TWITTER_TARGET_USERS=user1,user2,user3  # Comma-separated list of users to interact with
TWITTER_ENABLE_POST_GENERATION=true    # Set to false to disable automated posting
TWITTER_POST_INTERVAL_MIN=1            # Minimum minutes between posts
TWITTER_POST_INTERVAL_MAX=2            # Maximum minutes between posts

Important: Ensure you use the correct absolute path for PGLITE_DATA_DIR. Never commit your .env file containing real credentials to version control. Add .env to your .gitignore file to prevent accidental exposure.

Step 4: Start Your Project

Navigate to your project directory in the terminal and start the agent:

# On Windows:
# cd path\to\my-twitter-agent
# elizaos start

# On Mac/Linux:
cd my-twitter-agent
elizaos start

Your Twitter agent is now running and will operate based on the configured environment variables and character definition.

Customizing Your Twitter Agent's Personality

You can customize the agent's behavior, including its communication style, content focus, and interaction rules.

Open your project's src/index.ts file in a code editor. Locate the character definition:

export const character: Character = {
  name: 'Eliza', // Define the agent's name
  plugins: [
    // List of enabled plugins
    '@elizaos/plugin-sql',
    ...(process.env.OPENAI_API_KEY ? ['@elizaos/plugin-openai'] : []),
    ...(process.env.TWITTER_USERNAME ? ['@elizaos/plugin-twitter'] : []),
    // Add other plugins as needed
  ],
  settings: {
    secrets: {}, // Placeholder for plugin-specific secrets
  },
  system:
    'Define the core instructions for the agent's behavior on Twitter. Example: Be informative, share tech news, and answer AI-related questions.',
  bio: [
    'List personality traits or behavioral guidelines.',
    'Each string represents a distinct characteristic.',
    'Example: Focuses on AI developments.',
    'Example: Maintains a neutral and professional tone.',
    'Example: Includes links to sources in relevant posts.',
    'Example: Avoids engaging in controversial topics.',
  ],
  // Additional configuration options may exist below
};

Modify the following fields to tailor the agent's persona:

1. name: The agent's display name.
2. system: The primary prompt defining the agent's core function and behavior.
3. bio: A list of strings detailing specific personality traits, knowledge domains, or interaction rules.

After modifying the character definition, save the file and restart the agent using the development mode for faster iteration:

elizaos dev

Advanced Usage

For more fine-grained control, consider adjusting the following:

• Post Scheduling: Modify TWITTER_POST_INTERVAL_MIN and TWITTER_POST_INTERVAL_MAX in the .env file to control posting frequency.
• Topical Focus: Refine the system prompt in src/index.ts to guide the agent towards specific subjects.
• Interaction Logic: Update the bio list in src/index.ts to define how the agent responds and interacts.
• Targeted Interaction: Change the TWITTER_TARGET_USERS list in the .env file to specify which accounts the agent should primarily interact with.

Advanced: Server Deployment with Twitter Cookies

When deploying the agent to a server environment, standard username/password authentication might be insufficient due to Twitter's security measures. Using authentication cookies is often necessary.

Obtaining Twitter Cookies for Server Deployment

Follow these steps to extract the required cookies. Note that Twitter may flag logins from unfamiliar IP addresses, so performing these steps from an IP similar to your server's (e.g., via VPN) is recommended if possible.

1. Log in to Twitter via a web browser (Chrome, Firefox, Edge).

2. Open Developer Tools (usually F12).

3. Navigate to the Network tab.

4. Refresh the Twitter page (F5).

5. In the list of network requests, click on any request made to twitter.com.

6. Go to the Cookies tab within the request details.

7. Locate and copy the values for the following cookies:

   • auth_token
   • ct0
   • guest_id (This might be under a different domain like x.com or twitter.com depending on browser/timing).

8. Add the copied values to your .env file:

# Twitter Cookies for Server Deployment
TWITTER_COOKIES_AUTH_TOKEN=your_auth_token_here
TWITTER_COOKIES_CT0=your_ct0_token_here
TWITTER_COOKIES_GUEST_ID=your_guest_id_here

With these variables set, the agent running on your server should be able to authenticate using cookies.

Security Note: Remember to store your .env file securely and never commit it to version control, especially when containing sensitive cookie data. Ensure .env is listed in your .gitignore file.