LobeHub tutorial: create and schedule an AI agent

This Persian-language tutorial walks through LobeHub's hosted app, from account sign-in and language selection to creating an AI agent with a recurring task. The speaker mentions a Windows download, but the walkthrough uses the web interface. It does not demonstrate self-hosted deployment or running models locally.

The setup includes connecting Google Sheets and selecting preset agents for development work, such as code review and debugging. The speaker describes using connected services to record prices or manage appointments, including Google Calendar for bookings and Supabase for stored records. These are proposed workflows rather than completed demonstrations.

The main example is a custom walking guide for Tehran. The presenter writes its instructions, reports that web browsing is enabled, assigns it to a task, and sets a weekly schedule for Thursday at 10 a.m. The walkthrough also covers task priority and attachments. Creating the schedule does not establish that later runs completed successfully.

The speaker then introduces community agents, MCP tools and memory settings for personal context and preferences. Model usage is a practical limitation: the account shown has a token allowance, and the presenter warns that flagship models can exhaust it quickly. He recommends using lighter models while testing a workflow.