mobius-studio-mcp
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| SWEIPE_USER | Yes | The WordPress username with an Application Password. | |
| SWEIPE_SITE_URL | Yes | The URL of the WordPress site (e.g., https://example.com). | |
| SWEIPE_NAMESPACE | No | Optional override for the REST namespace (e.g., 'sweipe/v1' or 'flatmobile/v1'). If not provided, the server auto-detects. | |
| SWEIPE_APP_PASSWORD | Yes | The WordPress Application Password (spaces allowed). | |
| SWEIPE_IMPORT_BUDGET | No | Seconds the server spends per import step (default 25, max 55). Lower if your host kills long requests. | 25 |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": true
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| sweipe_statusA | What this WordPress install is (theme, plugin, Elementor, WooCommerce), whether the licence is active, whether AI is enabled and configured, and where the AI site plan stands (planned, built, reviewed). Call this first. |
| sweipe_list_demosA | The ready-made demo sites this licence can import (slug, name, description, category, required plugins). Use sweipe_import_demo with a slug. |
| sweipe_import_demoA | Import one of the demo packages into the site and wait for it to finish (pages, menus, images, settings; products when WooCommerce is active). Existing user pages are never overwritten. Takes one to a few minutes. |
| sweipe_import_statusA | Progress of a running or finished import by its import_id. |
| sweipe_ai_pingA | Round trip to the Sweipe AI service: confirms the licence is accepted and returns the credit balance (included monthly credits, used, left, reset date). |
| sweipe_ai_libraryA | The templates a site can be planned from (slug, name, what the design looks like, required plugins). Pass a slug as |
| sweipe_ai_get_planB | The stored site plan (pages and their sections), the brief it came from, and when it was built and reviewed. |
| sweipe_ai_planA | Turn a one-paragraph business brief into a site plan: pages, and for each page an ordered list of sections chosen from the template library. Costs 1 credit. Review the plan (edit page titles, slugs, add or drop sections) before sweipe_ai_build. |
| sweipe_ai_buildA | Assemble the plan into pages with copy written from the brief and stock photos, then import it into the site and wait for the import to finish. Costs 4 credits. Pass |
| sweipe_ai_reviewA | Render every built page, inspect it (empty sections, broken images, missing pages), take phone and desktop screenshots, and have the service judge the result against the brief. Returns a score out of 10, a summary, the changes made to the plan and whether the plan changed. If |
| sweipe_site_from_briefA | The whole loop in one call: plan from the brief, build and import, full review, and if the review changed the plan, rebuild and run a fix review. About 12 credits and 3–8 minutes. Use the individual tools when the plan should be edited by hand first. |
| sweipe_ai_set_enabledA | Switch the AI features of the plugin on or off for this site (Sweipe → AI). |
| sweipe_brief_pagesA | Every page, post and the site identity block (id 0: title, tagline, menus) with how many text items each holds, whether a draft is waiting, and whether AI copy has already been applied. Use the ids with sweipe_brief_generate / apply / undo. |
| sweipe_brief_pickB | Ask the service which demo packages fit the business brief, with a reason each. Free. Also stores the brief for later copy generation. |
| sweipe_brief_generateA | Rewrite every text on one page (or the site identity, post_id 0) for the business in the brief, keeping the layout. Returns the draft rows (id, widget, old, new) and stores them; nothing changes on the site until sweipe_brief_apply. 1 credit per page. |
| sweipe_brief_applyA | Write the stored drafts into the pages, keeping a backup for sweipe_brief_undo. Pass |
| sweipe_brief_undoA | Restore the pre-apply backup of the given pages (all backed-up pages when post_ids is empty). |
| wp_get_settingsA | Core WordPress settings (site title, tagline, timezone, language, front page) via wp/v2/settings. |
| wp_update_settingsA | Change core WordPress settings: title, description (tagline), timezone (e.g. "Europe/Istanbul"), language (e.g. "tr_TR"), show_on_front ("page"|"posts"), page_on_front, page_for_posts. |
| wp_list_pagesA | Published and draft pages with id, title, slug, link and status, for picking front pages and checking an import. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 20 tools
Most tools target distinct actions or resources: status checks, demo import, AI planning, and brief copy are clearly separated. A few boundaries could still trip an agent, such as sweipe_status vs sweipe_import_status or sweipe_list_demos vs sweipe_ai_library, but the descriptions provide enough to disambiguate.
The names consistently use snake_case and a sweipe_ prefix, with wp_ used for the WordPress core tools. There are minor structural deviations like sweipe_site_from_brief and sweipe_brief_pages where the noun/phrase form replaces the usual verb_noun pattern, but the overall scheme is predictable.
Twenty tools is on the heavy side for a single server, and the 16-25 range makes it feel sprawling. That said, the functions are grouped into distinct workflows (demos, AI site building, brief copy), so the count is defensible even if it could be consolidated.
The server covers the full Sweipe lifecycle: licensing/status, demo import, AI plan/build/review, one-shot brief-to-site, and the copy brief workflow with undo. Minor gaps exist, such as no direct arbitrary page editing or delete/cancel operations, but the core workflows are complete and have no dead ends.