Skip to main content
AI Chats are where you interact with Shinkai Web agents, send instructions, use tools and run workflows, all from your browser. This guide explains how to start chats, use tools, personalize prompts, and make the most of agent conversations. A quick walkthrough of sending your first message, switching agents,and enabling tools. Starting a Chat To start a chat in Shinkai Web you have 2 optiones: a)
  1. Go to My Agents or Explore Agents in the sidebar  
  2. Choose an agent from the gallery (or one you created)  
  3. Start chat type your message in the input bar
You can start as many chats as you want — each one is saved automatically. b)
  1. Go to + New Chat in the sidebar  
  2. Start chat type your message in the input bar
You can start as many chats as you want — each one is saved automatically. Switching Agents In any conversation, you can switch to another agent:
  • Click the agent name at the down of the chat  
  • Choose another agent from your list  
  • The new agent will instantly continue from the same context.
Multi-Agent Conversation Flow If enabled, agents can collaborate within a chat:
  • Agent A does research  
  • Agent B creates content  
  • Agent C edits or verifies  
Shinkai Web will show which agent performed each step. Chat Actions Inside each chat you can:
  • regenerate a response  
  • copy text  
  • duplicate the chat  
  • save a chat as a preset  
  • export content (when available)  
  • pin important messages
Chats are autosaved and can be resumed anytime. Best Practices

​
Be specific  

Clear instructions = better agent performance.

​
Provide data  

The more relevant files or context you add, the better the output.

​
Try different agents  

Writer, researcher, coder, creator — each has different strengths.

​
What’s Next?

After learning to run chats, explore:
  • Creating Agents on Web-Build your own agents with instructions, tools, and prompts.
  • Agent Collaboration -Run multi-agent workflows directly from the browser.
  • x402 Payments on Web-Use credits to run advanced agents and higher-volume tasks.