Weavz
Add Weavz to your AI client

Try Weavz from the AI client you already use.

Add one Weavz connector URL, sign in or create an account inside the authorization flow, choose a workspace, and let the agent use approved app actions with scoped credentials, files, state, approvals, and audit.

Try in Claude

You can start from the connector flow first; Weavz creates or signs in the account during authorization.

Connector details

Connect ChatGPT, Claude, or a coding agent to a Weavz workspace. The workspace decides which apps, credentials, files, state, browser tools, and approvals the agent can use.

ChatGPT
Claude
Coding agents
Weavz workspace
scoped tools, users, approvals, state, and audit
Connected apps
Connector name
Weavz Agent Workspace
Connector URL
https://platform.weavz.io/mcp/weavz
No Weavz account required before setup; the connector signs the user in during authorization.
One hosted MCP URL works in Claude, ChatGPT, and remote-MCP clients.
Access is scoped to one workspace and one end user after OAuth.
Connected app accounts stay behind Weavz approvals, aliases, state, files, and audit.
Setup paths

Start from the client in front of you.

Claude

Open the prefilled connector setup, confirm Weavz Agent Workspace, and complete the Weavz authorization flow. You can create the account and workspace during setup.

Try in Claude

ChatGPT

ChatGPT does not currently provide a reliable prefilled custom connector link. Use the fields below when creating a connector in ChatGPT settings.

Coding agents

Remote and local coding agents can connect through MCP. Add the hosted Weavz URL to Codex, Claude Code, Cursor, OpenCode, VS Code, Gemini CLI, or any HTTP MCP client.

Set up in coding agents
What happens next

The connector is just the door. The workspace carries the control.

Choose the workspace

Weavz scopes the connector token to one workspace and one end user.

Connect apps as yourself

Gmail, Slack, GitHub, Sheets, CRM, files, state, browser, and other aliases stay tied to the right user.

Review risky actions

Human Gates can pause external sends, shared writes, exports, spend, and customer-facing changes.

Coding agents

Use the same remote MCP resource from code-first clients.

Remote and local coding agents can connect to Weavz over MCP. Add the stable connector URL, complete Weavz sign-in, and the agent gets governed access to approved tools, files, state, and Human Gates through scoped app connections. Builders creating custom harnesses can provision workspace-specific MCP servers and per-user tokens from the API or SDKs.

Claude Code
Add the hosted MCP resource from the Claude Code terminal.
claude mcp add --transport http weavz https://platform.weavz.io/mcp/weavz
Codex CLI
Register Weavz and complete the MCP login flow.
codex mcp add weavz --url https://platform.weavz.io/mcp/weavz
codex mcp login weavz
Cursor / generic MCP config
Paste the remote MCP server into Cursor or another HTTP MCP config.
{
  "mcpServers": {
    "weavz": {
      "type": "http",
      "url": "https://platform.weavz.io/mcp/weavz"
    }
  }
}
Gemini CLI
Add Weavz as the remote MCP server your Gemini agent can call.
gemini mcp add --transport http weavz https://platform.weavz.io/mcp/weavz
OpenCode
Enable the Weavz MCP connector in your OpenCode config.
{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "weavz": {
      "type": "remote",
      "url": "https://platform.weavz.io/mcp/weavz",
      "enabled": true
    }
  }
}
VS Code / Copilot
Use the same hosted MCP URL where VS Code or Copilot expects MCP servers.
{
  "servers": {
    "weavz": {
      "type": "http",
      "url": "https://platform.weavz.io/mcp/weavz"
    }
  }
}
Running agents locally?

You can also use the Weavz CLI for terminal workflows: inspect workspace context, connect apps, search actions, run actions, and execute Code Mode from the same governed workspace.

CLI page
bash
brew install weavz/tap/weavz
weavz login

# or use npm
npm install -g @weavz-io/cli
weavz login

# no persistent install
npx -y @weavz-io/cli login