TweakBoard: Game Tuning Dashboard
... Formulas & Validation Derived fields built from expressions over other fields Validation rules with Info, Warning and Error levels ... One-click save for all modified assets and scenes
by swallow
Price History +
Every tuning value in one editable dashboard. Compare objects side by side, add formulas and validation rules, snapshot changes. No code needed. Built-in MCP server lets AI agents tune with you.
Every tuning value in one editable dashboard.
📋 Dashboard
- Collect fields from ScriptableObjects, prefabs and scene components into one window
- Edit your real game data, with full Undo
- Colored zones and cards with drag-and-drop arrangement
- Multiple layouts, saved as assets you can keep in version control
- Sidebar tree, search, tag filters and notes on any field
- No code required
➕ Fast Field Capture
- Add fields from the Inspector's right-click menu
- Record mode captures every field you click
- Drag and drop objects to add fields or whole cards
📊 Comparison Tables
- Every object of a type side by side in one table
- Automatic discovery across assets, prefabs and scenes
- Sorting, filtering and column visibility
- Editable tables for arrays of structs
🧮 Formulas & Validation
- Derived fields built from expressions over other fields
- Validation rules with Info, Warning and Error levels
- Rule presets for range, non-zero, field comparison and ratio
- Custom C# validators
- Warning counter and a filter for failing fields
🕓 Change Tracking
- Session change markers, with revert per field or for everything
- Named snapshots with diff and selective restore
- CSV export and import, with a preview before changes apply
- One-click save for all modified assets and scenes
🛡️ Robust References
- Fields keep working when assets are renamed or moved
- Edit values in scenes that aren't loaded
- Clear error reasons, plus smart relinking for renamed properties
🤖 AI Agents (MCP)
- Built-in MCP server for Claude Code, Codex, Kiro, Windsurf and GitHub Copilot
- AI agents can build dashboards, tune values and run validation
- Local only and off by default, with access levels
- One Undo step per agent action, an activity log and optional AI tagging