Real-World Use Cases

See how teams use xbot in practice.

Scenario: A 20-person engineering team wants AI help without everyone managing their own API keys.

Setup:

1. Deploy xbot in Server mode on a team server
2. Admin creates an LLM subscription via /setup
3. Connect a Feishu bot app
4. Add the bot to team group chats

Result: Anyone in the team @mentions the bot to:

• Generate code snippets and review PRs
• Look up documentation and answer technical questions
• Operate Feishu Docs and Bitable (create reports, update trackers)
• Run shell commands on the server (with permission control)

No individual API key management. The admin controls access via allow_from.

Scenario: A solo developer wants a powerful terminal AI assistant.

Setup:

1. Install in Standalone mode
2. Run xbot-cli in your project directory
3. The agent inherits your working directory and can read/write files

Result:

• Ask the agent to explore an unfamiliar codebase (explore SubAgent)
• Delegate focused tasks (code review, test writing) to SubAgents
• Run commands and debug issues interactively
• Use /context to manage token usage during long sessions

Scenario: You want the agent to run periodic checks and send alerts.

Setup:

You: "Every morning at 9 AM, check if the nightly CI passed.
If any job failed, summarize the errors and notify me."

Agent: *uses Cron tool to schedule*
Agent: "Done. I'll check CI status every morning at 9 AM and
notify you of any failures."

Result: The agent schedules itself via the Cron tool. In Server mode, the schedule survives restarts.

Scenario: You need multiple expert perspectives on a design decision.

Setup:

You: "Review this API design. Get input from a security expert,
a performance expert, and a UX expert, then synthesize."

Agent: *creates a Group Chat with three SubAgents*
Agent: "@security-expert what are the auth risks?"
Agent: "@performance-expert any bottlenecks?"
Agent: "@ux-expert is the API ergonomic?"
Agent: *synthesizes all three perspectives into a recommendation*

Result: The Group Chat Meeting Mode lets multiple specialized SubAgents debate and converge on a recommendation.

Scenario: Your team tracks work in Feishu Bitable and writes reports in Feishu Docs.

Setup: The agent has native Feishu tools:

You: "Read the project tracker Bitable, summarize the status of all
in-progress tasks, and create a weekly report Doc."

Agent: *uses feishu_bitable_list to read the tracker*
Agent: *uses feishu_docx_create to write the report*
Agent: "Done. I've created a weekly report Doc with the status summary."

Result: The agent reads Bitable records, processes data, and creates Feishu Docs — all through conversational commands.

Scenario: You want to offer AI chat access to team members who don’t use the terminal.

Setup:

1. Deploy Server mode with web.enabled: true
2. Enable invite-only mode for access control
3. Share invite links with team members

Result: Non-technical users access the agent through a browser-based chat UI with markdown rendering, code highlighting, and file uploads.

Note

Have a use case we should document? Open an issue on GitHub.

• Comparison — xbot vs other agents
• Tips & Tricks — power-user tips
• FAQ — common questions