PlayTonight

What is PlayTonight?

PlayTonight is an MCP server that enables querying your game library with a large language model for casual, intelligently calculated game suggestions.

How to use PlayTonight?

Configure the tool in Open-WebUI (configuration steps are not yet provided), then set the supplied system prompt. Use the /refine command to extract relevant keywords from the user’s conversational message, then pass those keywords to /recommend to get game suggestions. Always prioritize the user’s stated preferences over random selection.

Key features of PlayTonight

• Queries your personal game library for suggestions
• Uses /refine and /recommend commands for interaction
• Prioritizes user-stated preferences like genre, mood, or setting
• Returns playtime in minutes (should be converted to hours when displayed)
• Only suggests games returned by the tool server

Use cases of PlayTonight

• Decide what to play tonight based on current mood or genre interest
• Find a game from your library that matches a specific setting or theme
• Get a quick, intelligent recommendation when you can’t choose

FAQ from PlayTonight

What is PlayTonight?

PlayTonight is an MCP server that lets you query your game library with an LLM for casual, intelligent game recommendations.

What commands does PlayTonight use?

PlayTonight uses /refine to extract keywords from the user’s message and /recommend to return game suggestions based on those keywords.

How should playtime be presented to the user?

Playtime is returned in minutes; the LLM should convert it to hours when presenting the information to the user.

What should the LLM prioritize when making recommendations?

The LLM should prioritize the user’s stated preferences (e.g., genre, mood, setting). Random selection should be used only if no preferences are given.

Can the LLM suggest games not returned by PlayTonight?

No. The LLM must only suggest games that were returned by the tool server and must not invent responses.