How to Connect X (Twitter) to an AI Agent: Step by Step

· DoubleX

Quick answer

To connect X to an AI agent, open the Connectors screen in DoubleX Agentes, add the X connector, authorize it, then choose which agents can use it. The agent can search and read public posts and profiles. It is read-only: it cannot publish, reply, or send messages.

Many teams want an AI agent to keep an eye on what is said on X: mentions of a brand, a competitor's announcements, a public figure's posts. The X connector gives an agent that ability, and it works only with public content. This guide shows how to set it up, how to control it, and where its limits are.

What does the X connector let an agent do?#

The X connector lets an agent search and read public posts and public profiles on X. A connector is a system linked to the platform, and the agent uses the tools that system exposes. In DoubleX Agentes, the X connector is in the marketing category.

In practice, that means the agent can:

  • Search for public posts about a topic, product name, or keyword.
  • Read the posts of a public profile.
  • Read profile information that is publicly visible.
  • Summarize what it found and deliver it as a report, a spreadsheet, or a message.
TaskHow the agent helpsWhat it reads
Brand monitoringSearches for mentions and summarizes the tonePublic posts
Competitor trackingReads a competitor's recent posts and groups them by themePublic profile and posts
Market researchCollects posts about a topic and lists recurring pointsPublic posts
Content ideasReviews what an industry profile posted and suggests anglesPublic posts

What do you need before you start?#

You need a DoubleX Agentes account and at least one agent. You also need permission to add connectors to your workspace. Creating an account is free and does not require a card.

Before connecting, decide three things:

  • The goal. Monitoring mentions, researching competitors, and collecting market signals call for different instructions.
  • Which agent will use it. A marketing or research agent is a common choice.
  • Who should see the connection. A connection can be personal, for a team, or for the whole company.

If you have not created an agent yet, take the diagnostic first. It is a 5 to 10 minute conversation about your routine, and it can turn into a team of agents already loaded with your company's instructions. The Quickstart walks through that first setup.

How do you connect X on the Connectors screen?#

Open the Connectors screen, find X, and follow the authorization flow. Once the connection is active, the agents you allow can use it.

  1. Sign in to DoubleX Agentes and open Connectors.
  2. Find X in the list (marketing category) and start the connection.
  3. Complete the authorization when prompted.
  4. Choose the scope of the connection: personal, team, or company.
  5. Confirm that the connector appears as connected.

If you manage connectors through code, the Connectors API documentation describes how to handle them programmatically.

How do you choose which agents use X, and with what permissions?#

You choose this per agent. For each agent, the company decides which systems it uses and, within each system, which actions it can run. This keeps X available only where it is useful.

  1. Open Agents and select the agent that should use X.
  2. In the agent's connectors, enable X.
  3. Review the actions available for X and keep only the ones the agent needs.
  4. Add instructions about how to use it, such as which topics, profiles, and languages to cover.
  5. Save the agent.

Because the X connector reads public content only, the main control is deciding which agents can use it at all. Roles of owner, administrator, and member, plus teams, help you control who can change these settings.

How do you test the connection with a first request?#

Send a small, specific request in a conversation with the agent. Each request becomes a task on the Tasks screen, with the full history: the request, each tool the agent used, and the result.

Try one of these:

  • "Search X for public posts that mention our company name from the last few days and summarize the main themes."
  • "Read the recent public posts from this profile and list the announcements about new products."
  • "Find public posts about this topic and group them into questions, complaints, and praise."
  • "Put the posts you found into a table with the author, the text, and the link."

Open the task afterward to confirm the agent used the X connector. If the result is too broad, narrow the topic or the profile in your next request.

Once the first test works, you can ask for the same research on a schedule. The guide on researching X posts and profiles with AI has prompts and routines you can adapt. The Triggers documentation covers how scheduled routines work, which run with the permissions of the person who created them.

What does the X connector not do?#

The X connector is read-only for public content. It does not publish posts, reply, send direct messages, or access private accounts.

The connector doesThe connector does not
Search public postsPublish or schedule posts
Read public profilesReply, like, or follow
Feed research into reports and spreadsheetsSend direct messages
Work alongside other connectors you enableRead private or protected content

If you want the findings to go somewhere, pair X with another connector that the agent is allowed to use, such as a spreadsheet or a document tool. Those other actions follow their own permissions, and the agent asks for confirmation before sending, deleting, or changing anything in your systems.

Frequently asked questions#

Can the AI agent post on X for me?#

No. The X connector in DoubleX Agentes only searches and reads public posts and profiles. It does not publish, reply, or send messages.

Can the agent read private or protected accounts?#

No. The connector works with public content only. Anything that is not publicly visible on X is outside its reach.

Do I need to know how to code to connect X?#

No. You add the connector on the Connectors screen and describe what you need in your own words. APIs, a command line, and an MCP server exist for people who want to go further.

Does the content from my research train AI models?#

No. Your company's content is not used to train models, and requests go only to AI providers that do not use what they receive for training. You can also choose zero retention on the Models screen.

Can I run X research automatically every day?#

Yes. You can set up a routine so the agent runs at a set time, such as every morning, using the permissions of the person who created it. Every run is recorded in the task history.

Related guides

More X guides