For AI agents

Read the studio's services, explore examples and prepare drafts through structured tools.

When to use this studio

Use Humble Monkeys Studio for custom websites, apps, brand work, automation and AI interfaces, especially hospitality and tourism. Read the catalogue, compare fictional concepts, then prepare a project brief for human review. Scope and investment require a studio proposal. Tools only compute drafts and previews; they do not send messages, make bookings, publish, call anyone or change browser saves. Use sample data in demonstrations. Read /for-ai and /privacy for details.

For hospitality or tourism, start with website design and the hotel concept. The roofing and salon concepts show other business directions. All three are fictional studio examples.

Read any page

Every public page has a static HTML version and a Markdown version containing its main content, links, tables and form guidance. Request the page with Accept: text/markdown on humblemonkeys.com, or use its direct Markdown link. Missing pages return HTTP 404 with a recovery link in the requested format.

Connect an agent

The public, stateless MCP endpoint is https://humblemonkeys.com/mcp. It uses Streamable HTTP and supports protocol versions 2026-07-28, 2025-11-25, 2025-06-18 and 2025-03-26. No account or API key is needed for these draft and preview tools. Discover tools through tools/list and public page resources through resources/list.

Browser agents can use the same tools through WebMCP when the browser supports document.modelContext. Page forms also describe their actions and fields for WebMCP. Form submission keeps the browser's human review step. Standard links, labelled controls and visible results remain usable in other browsers.

Available tools

  • Preview a hotel reservation — hotel_quote_stay. Check fictional room inventory, capacity, nightly rates, extras, promo codes and a deposit. CAD amounts are cents, with illustrative 10% tax. No real availability, booking or payment.
  • Preview a hotel guest service — hotel_preview_request. Preview fictional room service, dining, spa, transport, activities, housekeeping, maintenance, access or late checkout. Read the hotel page for options and times. Nothing is requested externally.
  • Preview a hotel group quote — hotel_quote_event. Calculate a fictional hotel event estimate, checking venue capacity and adding catering, room nights and a deposit. CAD cents with illustrative 10% tax. No venue or room block held.
  • Preview hotel channel distribution — hotel_preview_channels. Preview sample rates and one shared inventory pool across four booking channels. Commissions are fictional estimates. No provider connection or synchronisation.
  • Calculate hotel reporting metrics — hotel_calculate_metrics. Calculate sample occupancy, ADR and RevPAR from aggregate room nights and CAD cents. Zero denominators return null. No live PMS or accounting data.
  • Ask the sample hotel concierge — hotel_concierge_answer. Read preset fictional hotel information in English, French or Spanish. Covers arrival, dining, spa, transport, access, local activities and reservations. No connected AI or staff message.
  • Preview a hotel operations task — hotel_preview_operation. Prepare a fictional housekeeping, maintenance, guest message, handover or purchase task with owner, priority and checklist. No staff action, save, message or purchase.
  • Explore studio services and examples — studio_catalog. Read the service catalogue, fictional website concepts, demo limits, contact details and all page URLs. Optional plain-text search.
  • Prepare a project brief — studio_prepare_brief. Create an unpriced custom project brief for human review. No enquiry is sent and no agreement is made.
  • Build a sample itinerary — studio_preview_stay. Build a fictional retreat itinerary using the visible Stay Planner logic. CAD activity estimates exclude meals, lodging, taxes and transport. No live availability, weather or bookings.
  • Draft a review reply — studio_draft_review. Create a local template-based review reply for human editing. No connected AI, verified facts or automatic publishing. Include a public contact only with explicit permission.
  • Generate a sample content calendar — studio_plan_content. Create editable template-based content ideas with the Content Calendar logic. All captions require factual review; nothing is published or scheduled externally.
  • Calculate a sample proposal — studio_calculate_proposal. Validate and calculate your supplied sample proposal using exact cents arithmetic. This is not a Humble Monkeys quote. No document is sent or accepted.
  • Preview a lead record — studio_validate_lead. Validate a sample Lead Desk record and return a normalized record and pipeline summary. It does not create a lead, save data or contact anyone.
  • Prepare an onboarding checklist — studio_prepare_onboarding. Return a fictional Client Launchpad template with task dependencies and optional project profile. Nothing is shared, uploaded or saved.
  • Compare dashboard periods — studio_compare_periods. Validate two non-overlapping sample reporting periods and calculate Pulse Board metrics. Counts are aggregate period totals, not a tracked cohort. Zero denominators return null with an explanation.
  • Preview a fictional business request — studio_preview_concept. Preview a stay enquiry, roof estimate request or salon appointment in the fictional studio concepts. No availability is checked; nothing is booked or sent.
  • Try a sample receptionist conversation — studio_receptionist_turn. Run one turn of the browser receptionist demo. Carry the returned state into the next call. Use sample details. No live call, transfer, callback or message is made.
  • Interpret a canvas command — studio_canvas_command. Preview the preset canvas parser: add a table/image/video, write a headline, remove a type, or clear. Returns the interpreted action without changing or saving a canvas.
  • Read the sample enquiry workflow — studio_workflow_preview. Return the unpriced sample enquiry-to-brief workflow and its human review points. Approval in this demonstration is not a project agreement.

Prepare a real project enquiry

Read the catalogue, choose service identifiers and call studio_prepare_brief with the project name, goals and audience. Show the brief to the person requesting the work. The contact form sends a real enquiry after human review and a security check. A successful server response confirms submission to the studio; preparing a brief does not. Scope, deliverables, responsibilities and investment are agreed in a separate proposal.

Data and completion

Browser demos work locally. Calling the remote MCP endpoint transmits the fields you provide to Cloudflare for that calculation; the application does not save the inputs or results. Your agent provider may handle conversation data according to its own settings. Use fictional data when trying demonstrations. Tools do not send email, place calls, make reservations, publish content, upload files, sign agreements or change local browser saves.

Invalid inputs return structured field errors. Retry after correcting those fields. Successful previews include explicit demo status and a link to the corresponding workspace. Follow the human review steps before copying, sending or using a draft. Read the privacy details.