Skip to content
webauditlabs docs

AI agents and MCP

CLI

Section titled “CLI”

Agents can run the CLI directly and read the JSON result:

Terminal window
npm run webauditlabs -- run webqa.json --json

The result contains rule identifiers, severities, pages, viewports, elements and screenshot paths. After a fix, the agent runs the same command again: exit code 0 confirms that no findings remain.

Local MCP server

Section titled “Local MCP server”

webqa mcp starts an MCP server over stdio for Claude Code, Cursor and other MCP clients. It runs the local engine and needs no account.

{
  "mcpServers": {
    "webqa": { "command": "npx", "args": ["webqa-platform", "mcp"] }
  }
}

Proposed tools: audit_run, audit_status, findings_list, evidence_get, audit_rerun. The same safety guards as the CLI apply; private addresses must be allowed explicitly.

Remote MCP and REST API (webauditlabs Cloud)

Section titled “Remote MCP and REST API (webauditlabs Cloud)”

For cloud projects: projects_list, checks_list, checks_run, runs_get, findings_list, evidence_get, runs_compare, checks_propose, with scoped API tokens (projects:read, checks:run, findings:read). Available on every plan, including Free. Both MCP servers return the same findings as the CLI.