agent-builder

To build AI apps that interact with the OpenClaw agent, you use this framework to define a manifest, create chat sessions, and expose app tools for the agent to invoke. It requires permissions to access chat, memory, and tools within the agent ecosystem.

Streaming chat in DenchClaw apps works by creating and managing chat sessions with the OpenClaw agent, allowing your application to send messages and receive streamed responses in real-time for interactive automation scenarios.

You expose your app's functions as tools for the AI agent to invoke by configuring app-as-tool exposure within the defined manifest, granting the agent permission to trigger your app's internal scripts and capabilities during a chat session.

Yes, you can read and leverage agent memory for persistent context in your AI application. This feature allows your DenchClaw app to access historical data and maintain continuous session management across real-world automation scenarios.

Yes, creating a DenchClaw app requires a clearly defined manifest structure. The manifest provides a blueprint for app layout and basics, and you can optionally include scripts, references, and assets to extend the app's functionality.

Building AI apps with session management and agent memory requires explicit permissions to access chat, memory, and tools within the agent ecosystem. App functionality is constrained by the defined manifest and any optional scripts, references, and assets you configure.