Read-only Steam wishlist, sales and refund data for game developers, from Steamworks. Ask Claude or ChatGPT about wishlist adds and deletes, wishlist-to-purchase conversion, sales and refunds by country, and how Steam events like Next Fest moved your numbers. Connects to your own Steamworks Financial API key; without one it answers from a sample studio so you can try it. Sign-in with OAuth; it never writes to Steam.
Scored across 9 tools
Most tools target a clearly distinct slice: account metadata, game list, event calendar, sales, wishlist activity, breakdown, conversion, per-title comparison, and event comparison. The main soft overlaps are compare_titles versus get_sales (which can also be scoped to the whole account) and get_event_impact, which re-packages the same sales/wishlist metrics over a comparison window; descriptions do enough to steer selection.
Names are uniformly snake_case with a leading verb (get_, list_, compare_), giving a predictable read. The only minor deviation is vocabulary rather than structure: 'titles' in compare_titles versus 'games' in list_games, which slightly muddies the resource naming.
Nine tools is well-scoped for a read-only Steam analytics server, with each endpoint covering a distinct axis (time series, breakdown, conversion, comparison, reference data). Nothing appears redundant or padding.
The wishlist and sales lifecycle is well covered: activity over time, geographic/platform origin, conversion to purchases, revenue/refunds, event impact, and per-title comparison, plus the reference lists needed to drive them. Gaps are minor: no traffic/impression-to-wishlist funnel data, no way to export or pull wishlist targets/forecasts, and no summary of total outstanding wishlists across the account in a single call.
One row per game on the account for a period: wishlist adds, net wishlist change, outstanding wishlists, units, refunds and net revenue. Useful for publishers with several titles.
Inputs: end_date, start_date.
What this connection reads: sample or live data, the studio, how far the data goes, and when it was last updated.
Inputs: no named parameters.
Compare daily wishlist adds, wishlist purchases, units and net revenue during an event (a Steam event id from list_steam_events, or your own start and end dates such as a launch or your own discount) against the days just before it.
Inputs: appid (required), end_date, event_id, start_date, +1 more.
Units sold, units returned (refunds), gross, tax and net revenue in USD, for one game or the whole account. Group by day, week, month or total, and break down by country, discount level, package or app. Refund rate per row. Days are Pacific time, as Valve reports them.
Inputs: appid, limit, country, end_date, +3 more.
Wishlist adds, deletes, purchases and gifts for one game, grouped by day, week (Monday start) or month, with the number of outstanding wishlists at the end of each period. Optional single-country filter. Days are GMT, as Valve reports them.
Inputs: appid (required), country, end_date, group_by, +1 more.
Where wishlists come from for one game: by country (adds, deletes, purchases, share of adds, and outstanding wishlists per country at end_date) or by platform (Windows, Mac, Linux share of adds).
Inputs: appid (required), by, limit, end_date, +1 more.
How many outstanding wishlists turned into purchases or gifts in a period (for example launch week or a sale), and what share of units sold came from wishlists. Optional per-country split. For unreleased games it returns the outstanding wishlist count instead.
Inputs: appid (required), end_date, by_country, start_date.
List the games, DLC and soundtracks on the Steam partner account, with app ids, release dates (or coming soon), prices and how far back data goes. Call this first to get app ids.
Inputs: no named parameters.
Steam Next Fest and seasonal sale dates from Valve's published calendar, with ids for get_event_impact.
Inputs: no named parameters.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Add one secure layer between your agents and this server.