Wheelhouse
Roadmap
Feedback
Changelog
←
Back to changelog
Powered by Canny

new

Wheelhouse for AI Agents: A New Plugin Library for a Faster MCP

Plugin Library
TL;DR →
  • The
    Wheelhouse Plugin Library
    is live!
  • Our plugin library contains a
    growing set of Revenue Management skills
    that you can leverage (STLY pacing, price-change attribution, portfolio leaderboards, and more)
  • These can be
    installed in one step
    in
    Claude, Cursor, Codex, Grok Build, ChatGPT,
    etc.
  • This significantly
    improves the efficiency of our MCP
    , in some cases reducing the resources (tokens) required for common requests
    by 96%
    (i.e. pulling all reservations from WH is hugely improved)
  • Think of this as
    version control
    for our APIs & MCP.
  • We updated our RM API docs, to
    better serve agents & LLMs
    .
  • 📽️ Loom walkthrough for getting started today.
🔌 The Wheelhouse Plugin Library
Today, we’re sharing the Wheelhouse Plugin Library — a public, growing collection of Revenue Management skills you can connect straight into your AI client.
If you’re already using the Wheelhouse MCP, you know it gives your agent direct access to your portfolio — listings, pricing, reservations, preferences, and more.
The Plugin Library goes a step further: it packages that MCP connection together with the actual workflows our Revenue Management team builds, so your agents can do real analysis, faster & easier.
A few of the skills already in the library:
  • STLY pacing
    — on-the-books pacing for any listing or segment, over any date range, compared to the same time last year
  • Price-change attribution
    — checks whether a recent rate or preference change actually correlated with bookings
  • Future rate overpricing
    — flags months where your posted rates look out of line with what actually transacted last year
  • Custom rate attribution
    — a fast check on whether a custom rate you set actually got booked
  • Portfolio leaderboards
    — ranks your whole portfolio by pacing, market position, expiring inventory, and more
You don’t have to name a skill or remember which one to use — ask in plain English and your agent picks the right one:
  • “How is this listing pacing versus last year?” → STLY pacing
  • “Did that rate change actually drive bookings?” → Price-change attribution
  • “Are next season’s rates too high?” → Future rate overpricing
  • “Who needs attention across my portfolio?” → Portfolio leaderboard
🥳 Pro Tip!
Once you’re connected, ask for STLY pacing on any listing — or a whole segment, across any date range you choose — and get back the same pacing logic our own team relies on, already vetted by Wheelhouse.
🚄 A Faster, More Efficient MCP
Under the hood, we’ve also made a real efficiency improvement to how the Wheelhouse MCP handles complex, data-heavy requests.
Previously, when your agent needed to work through a lot of data, it worked one step at a time: call a tool, wait for the full result to come back through the model, decide what to do next, call the next tool. Every one of those round trips costs tokens and time.
Now, for multi-step, data-heavy requests, your agent can write a short piece of code that calls the tools it needs, loops through the results, and filters everything down — all before anything comes back into its context. Instead of pulling every reservation or every day’s pricing data through the model just to produce a summary, only the answer you actually asked for makes the trip.
As an example, this enabled us to reduce pulling reservations by 96%!
🥳 Pro Tip!
Try asking your agent to pull reservations across your entire market or portfolio in one go — this request that used to hit a wall and now runs end-to-end.
Getting Started:
Connecting is a one-step install, and our Plugin Library is supported everywhere: Claude Code, Cursor, Codex, Grok Build, and ChatGPT (Enterprise/Work, via admin-managed import) can all run the library.
And, the plugin library updates automatically as we ship new changes — there is nothing to reinstall/update/etc.
Here’s how easy it is in Claude Code, as an example — two lines, pointing at one repo:
/plugin marketplace add pricemethod/wheelhouse-plugin
/plugin install wheelhouse-plugin@wheelhouse-plugin
See a full walkthrough of connecting it in Claude here:
📽️ Loom walkthrough
🔗 Links:
Instructions for Cursor, Codex, Grok Build, and ChatGPT are linked at the bottom of this post.
grok plugin marketplace add pricemethod/wheelhouse-plugin
grok plugin install pricemethod/wheelhouse-plugin --trust
And, for more, you can always review our API docs: