SEOLZ speaks MCP
Fix it in your own editor. Your AI agent does the work.
Connect Claude Code, Cursor, Windsurf, or VS Code to SEOLZ over the Model Context Protocol and your coding agent works a site's fix worklist right in your repo — each fix with the exact failing element and platform-specific steps. You apply it; SEOLZ verifies it. We never touch your code.
AI coding access is a Pro and Agency capability.
What is MCP?
The Model Context Protocol is an open standard for connecting AI applications to external tools and data. Your AI app — the client (Claude Code, Cursor, Windsurf, VS Code, Claude Desktop) — talks to a server that exposes a set of tools through one uniform interface. Any MCP client can use any MCP server, no custom glue.
SEOLZ runs a secured MCP serverthat exposes your site's fix worklist as tools. Once your agent is connected, it pulls the next fix, applies it in your repo, and asks us to verify it — the same worklist you see in the dashboard, driven from your terminal or IDE.
The same worklist SEOLZ gives you — handed to your agent.
It's the exact loop we used to fix our own site across dozens of findings in a few sessions: pull, apply, verify.
Pull the worklist
Build — or resume — the prioritized fix list for your site: batched by fix surface (a template edit that clears dozens of pages vs. a one-off), severity-ordered, resumable across sessions and context resets.
Get each fix with real evidence
One fix at a time with the ACTUAL failing element on each page — the selector + colors for accessibility, the over-long value for SEO, the missing schema for AEO — plus platform-specific steps, a code example, and a ready-to-paste prompt.
Apply it in your own editor
Your agent edits your repo; you mark each fix done or skip it with a reason. SEOLZ never touches your code.
Verify after you deploy
Machine-verify a batch — title/meta/schema/canonical/security headers re-checked instantly; accessibility on the next full scan. Proof it landed, not just a checkbox.
Setup lives in your dashboard.
On Pro or Agency, open your site → Settings → Fix with your AI agent, generate a per-site token, and copy the ready-made setup for your client. You get a one-command connect for Claude Code and a copy-paste config for Cursor, Windsurf, and VS Code — with your token already filled in.
Per-site tokens
Each token is locked to one site and shown once. Generate one per site; revoke anytime.
Works with your client
Claude Code, Cursor, Windsurf, VS Code (Copilot agent), Claude Desktop — any MCP client over HTTP.
Built for developers
Stop copy-pasting findings into your editor. Your agent reads the worklist, opens the right files, and makes the change with the platform-specific steps already attached.
Built for agencies
One token per client site — locked to that site, never able to touch another. Point your agent at each site's worklist and clear findings across a portfolio without leaving your workflow.
Safe by design
SEOLZ never edits your repository — your agent does, in your own editor. Tokens are per-site, shown once, and revocable from Settings. Fixes only count once the next scan confirms them.
Frequently asked questions
What is MCP (Model Context Protocol)?+−
MCP is an open standard that lets an AI application (the client — Claude Code, Cursor, Windsurf, VS Code, Claude Desktop) connect to external tools and data through one uniform interface exposed by a server. SEOLZ runs a secured MCP server that exposes your site's fix worklist as tools your agent can call.
How do I connect my AI agent?+−
On the Pro or Agency plan, open your site in the SEOLZ dashboard and go to Settings → Fix with your AI agent. Generate a per-site token and copy the ready-made setup for your client (Claude Code, Cursor, Windsurf, or VS Code). The connection endpoint and step-by-step config are provided there.
Which AI clients can connect?+−
Any MCP client that speaks streamable HTTP — including Claude Code, Cursor, Windsurf, VS Code (Copilot agent mode) and Claude Desktop.
Does SEOLZ edit my code?+−
No. SEOLZ never touches your repository. Your agent applies every change in your own editor; the token only reads the worklist and records what you marked fixed. Each token is locked to one site and can be revoked anytime.
Point your agent at your site.
AI-agent access comes with Pro and Agency. Pick a plan, generate a per-site MCP token, and let your agent start clearing the worklist. We've got you.