X (Twitter)

XTools allows an Agent to interact with X, providing functionality for posting, messaging, and searching tweets.

Prerequisites

Install the required library:

1uv pip install tweepy

InfoTweepy is a Python library for interacting with the X API.

Setup

Create X Developer Account
- Visit developer.x.com and apply for developer access
- Create a new project and app in your developer portal
Generate API Credentials
- Navigate to your app's "Keys and tokens" section
- Generate and copy these credentials:
  - API Key & Secret
  - Bearer Token
  - Access Token & Secret

Configure Environment

1export X_CONSUMER_KEY=your_api_key
2export X_CONSUMER_SECRET=your_api_secret
3export X_ACCESS_TOKEN=your_access_token
4export X_ACCESS_TOKEN_SECRET=your_access_token_secret
5export X_BEARER_TOKEN=your_bearer_token

Example

1from kern.agent import Agent
2from kern.tools.x import XTools
3
4# Initialize the X toolkit
5x_tools = XTools(
6    wait_on_rate_limit=True # Retry when rate limits are reached
7)
8
9# Create an agent equipped with X toolkit
10agent = Agent(
11    instructions=[
12        "Use X tools to interact as the authorized user",
13        "Generate appropriate content when asked to create posts",
14        "Only post content when explicitly instructed",
15        "Respect X's usage policies and rate limits",
16    ],
17    tools=[x_tools],
18    )
19
20# Search for posts
21agent.print_response("Search for recent posts about AI agents", markdown=True)
22
23# Create and post a tweet
24agent.print_response("Create a post about AI ethics", markdown=True)
25
26# Get user timeline
27agent.print_response("Get my timeline", markdown=True)
28
29# Reply to a post
30agent.print_response(
31    "Can you reply to this [post ID] post as a general message as to how great this project is: https://x.com/AgnoAgi",
32    markdown=True,
33)
34
35# Get information about a user
36agent.print_response("Can you retrieve information about this user https://x.com/AgnoAgi ", markdown=True)
37
38# Send a direct message
39agent.print_response(
40    "Send direct message to the user @AgnoAgi telling them I want to learn more about them and a link to their community.",
41    markdown=True,
42)
43
44# Get user profile
45agent.print_response("Get my X profile", markdown=True)

Note

Check out the Tweet Analysis Agent for a more advanced example.

Toolkit Params

Parameter	Type	Default	Description
`bearer_token`	`str`	`None`	Bearer token for authentication
`consumer_key`	`str`	`None`	Consumer key for authentication
`consumer_secret`	`str`	`None`	Consumer secret for authentication
`access_token`	`str`	`None`	Access token for authentication
`access_token_secret`	`str`	`None`	Access token secret for authentication
`include_post_metrics`	`bool`	`False`	Include post metrics (likes, retweets, etc.) in search results
`wait_on_rate_limit`	`bool`	`False`	Retry when rate limits are reached

Toolkit Functions

Function	Description
`create_post`	Creates and posts a new post
`reply_to_post`	Replies to an existing post
`send_dm`	Sends a direct message to a X user
`get_user_info`	Retrieves information about a X user
`get_home_timeline`	Gets the authenticated user's home timeline
`search_posts`	Searches for tweets

You can use include_tools or exclude_tools to modify the list of tools the agent has access to. Learn more about selecting tools.

Developer Resources

View Tools