Retrieve the user's owned Steam game library with playtime statistics. Use to understand what the user owns, identify backlog (unplayed), or check playtime for specific games. Sort by 'recent' to see current obsessions. Filter 'unplayed' for backlog candidates.

Games played in the last 2 weeks — the strongest signal for the user's current mood and taste. Call this before recommend_from_library to understand recent preferences.

Retrieve the user's Steam wishlist. Use this to find games the user has already flagged as interesting, then pass the AppIDs to itad-mcp's get_game_prices or get_historical_low to check for deals. This is the most targeted buy-signal workflow.

Full Steam Store metadata for a specific game: description, genres, categories, price, platforms, Metacritic score. Use when evaluating a specific candidate. Pair with get_app_reviews for Steam sentiment and opencritic-mcp for critic scores.

Steam user review sentiment for a game: overall verdict, positive %, and total review count. Use alongside opencritic-mcp critic scores for a complete quality picture.

Resolve a game title to a Steam AppID. Use when you know a game's name but need its AppID for other tools. Returns multiple matches ranked by relevance.

Find games in the user's library that match specified tags/mood. Ideal for backlog recommendations. The LLM should derive tags from the user's mood or preferences (e.g., 'relaxing' -> ['Casual', 'Puzzle', 'Atmospheric']). Returns ranked matches with tag overlap scores.

Find well-reviewed games on Steam matching specific tags that the user does NOT own. Returns AppIDs — pass them to itad-mcp for pricing and deal data. Use when looking for new purchases.