Use Vexly Outside the Dashboard

Vexly can be used from scripts, terminals, agent workflows, and MCP-compatible desktop tools. The golden-path entry points are:

• Public API for scripts, automations, and CI jobs
• Vexly CLI for terminal use and lightweight wrappers
• Remote MCP for tools that can call MCP servers over HTTP

Choose the Right Method

Use the Public API if you want:

• A simple HTTP integration
• Language-specific script examples
• CI or automation jobs
• Direct control over polling and run handling

Start with the Public API guide.

Use the CLI if you want:

• A terminal command for one-off prompts
• A simple wrapper for agent-driven workflows
• A local helper you can script around

Start with the CLI guide.

Use MCP if you want:

• Claude Code, Codex, Claude Desktop, Cursor, VS Code, or OpenClaw integration
• A standard tool surface that exposes vexly.run
• A remote endpoint for MCP-capable clients

Start with the MCP guide.

Authentication

Vexly supports three auth paths for these external surfaces:

• Personal access tokens for scripts and CI
• Device flow for terminals and CLI-based flows
• OAuth browser flow for browser-capable remote MCP clients

See Authentication for the recommended auth path for each client type.

Public Repo

Reference assets for the CLI, MCP setup, and agent skills live in:

• just-strings/vexly-mcp

That repo is the canonical public distribution point for:

• CLI helper code
• MCP setup examples
• Claude Code and Codex skill files
• API examples