CustomGPT for Openclaw

A powerful integration for managing CustomGPT projects and knowledge sources through automated workflows.

membranedev
v1.0.4
Apr 30, 2026
0
421
0

Install & Download

1. ClawHub CLI

The fastest way to install a skill directly from the registry.

npx clawhub@latest install customgpt

2. Manual Installation

Copy the skill folder to one of these locations

Global
~/.openclaw/skills/
Workspace
<project>/skills/

Priority: Workspace > Local > Bundled

3. Prompt Installation

Copy this prompt to OpenClaw to install it automatically.

Help me install customgpt using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).

Prefer to download?

Get the raw skill files in a ZIP archive.

What is CustomGPT?

CustomGPT allows users to create highly customized AI chatbots powered by their own proprietary data. By leveraging this integration within Openclaw Skills, developers and businesses can programmatically manage the entire lifecycle of their custom copilots, from ingesting diverse knowledge sources like PDFs and websites to initiating real-time chat sessions.

This skill acts as a bridge between your AI agent and the CustomGPT platform, utilizing the Membrane framework to handle complex authentication and credential refreshes. This ensures that your agent can interact with CustomGPT data securely and efficiently without manual intervention or hardcoded API keys.

CustomGPT Use Cases

  • Creating and updating custom AI agents using specific sitemaps or document uploads.
  • Automating the retrieval of conversation histories and user messages for analysis.
  • Managing diverse knowledge sources including Google Drive, Notion, and Microsoft Office files.
  • Integrating CustomGPT chat completions into existing OpenAI-compatible workflows.
  • Programmatically deleting or cleaning up outdated data sources and indexed pages.

How CustomGPT Works

  1. The AI agent initiates a connection to CustomGPT via the Membrane CLI to establish a secure authenticated session.
  2. Knowledge sources such as text files, PDFs, or URLs are identified and linked to specific CustomGPT projects.
  3. The agent uses natural language intent to search for specific actions like list-agents or create-conversation.
  4. Actions are executed through the Membrane proxy, which automatically injects required headers and handles API communication.
  5. The agent processes the output to either provide information to the user or trigger further automation steps.

CustomGPT Setup

First, install the Membrane CLI to manage your Openclaw Skills:

npm install -g @membranehq/cli@latest

Next, authenticate your environment by specifying your agent type:

membrane login --tenant --clientName=openclaw

Ensure a connection to the CustomGPT service:

membrane connection ensure "https://customgpt.ai/" --json

CustomGPT Data Schema & Taxonomy

The skill organizes data around Agents (Projects), Conversations, and Sources. Below is the primary entity structure:

Entity Description Key Metadata
Agent The core chatbot project id, name, persona, settings
Source Knowledge inputs sitemap URL, file path, type
Conversation Active chat instances conversationId, user profile
Page Specific indexed items pageId, source origin

CustomGPT Advanced Features

  • Multi-agent support by adjusting agentType during the login process.
  • Long-polling connection states with the --wait flag for automated headless setups.
  • Direct API proxy requests for accessing edge-case endpoints not covered by standard actions.
  • OpenAI-compatible chat completion format for drop-in replacement in legacy systems.
  • Natural language action discovery to find the most efficient tool for a specific task.

SKILL.md


Loading

Related Openclaw Skills

METADATA

Github Stars: 0
forks: 0

Featured*