Overview

SpiderPublish was built for AI agents from day one. Anything you can do in the dashboard, an agent can do too — create pages, write posts, build forms, apply themes, manage domains, and deploy — because every surface sits on the same API.

Why this matters

Most website tools bolt on an API as an afterthought. SpiderPublish treats agents as a primary user. That means you can:

  • Describe what you want and get it built. "Add a pricing page with three tiers and publish it" becomes real changes you can review.

  • Automate the repetitive parts of running a site — drafting posts, updating copy across pages, refreshing a sitemap.

  • Keep a human in the loop. Agents preview destructive actions and wait for your confirmation, so you are never surprised.

The ways agents connect

There are three ways an agent or developer can drive SpiderPublish, all backed by the same API and the same safety model:

  • MCP — the Model Context Protocol. Connect an MCP-capable agent (Claude, Cursor, Windsurf, and others) and it gains a set of SpiderPublish tools it can call directly. Start at MCP Setup.

  • CLI — the spideriq command-line tool, for scripts, terminals, and CI. See The CLI.

  • VS Code extension — pull your content into your editor, change it, and push it back, with native diffs. See VS Code Extension.

Built-in safety

Every agent surface enforces the same rules: changes are scoped to your workspace, and destructive actions (publish, delete, deploy) run through a preview, then confirm flow. An agent gets a preview and a confirmation token; nothing lands until that token is used. Read Safe Publishing before you let an agent run.

A self-describing API

Agents do not have to guess how SpiderPublish works. A built-in reference endpoint returns a compact description of the content types, block types, and recipes, so an agent can learn the system on the fly and then build correctly the first time.

When to use an agent (and when not to)

Reach for an agent when you want to describe an outcome and have it built or maintained — adding pages, drafting and updating posts, building forms, applying themes, and deploying — or when you want to automate repetitive site work in scripts or CI.

If you are making a quick, one-off edit you can see in front of you, the dashboard is faster. The agent surfaces shine for repeatable, multi-step, or programmatic work — not for a single hand edit.

Where to start

  1. Get a token.

  2. Connect via MCP or install the CLI.

  3. Read Safe Publishing.

  4. Ask your agent to build something.

Next steps