fix: README.yaml metadata for HF Spaces

README.md

@@ -1,67 +1,48 @@
-# LoomLens Live — Standalone Package
-
-**Location:** `signal-loom-loomlens/`  
-**Status:** Phase 1–3 built. Packaging for release TBD.
-
 ---
+title: LoomLensAI — Token Budget Calculator
+emoji: ⚡
+colorFrom: purple
+colorTo: purple
+sdk: static
+app_file: index.html
+pinned: false
+---
+
+# LoomLens AI™ — Token Budget Calculator
 
-## What's Built
+**One prompt. Every model. Every cost. Before you run.**
 
-| File | Phase | Description |
-|------|-------|-------------|
-| `loomlens-models.ts` | 1 | Model registry — 12 models across 5 tiers with live pricing |
-| `loomlens-engine.ts` | 1 | Token estimator + task detection + cost calculator + fit scorer |
-| `loomlens-panel.tsx` | 2 | React pre-flight panel with model cards + tip footer |
-| `loomlens-openclaw-plugin.ts` | 3 | OpenClaw plugin with `before_model_resolve` hook + routes |
+⚡ Find the best model to fit your prompt — free, 5 estimates/day.
 
-## Quick Start (Engine only — no UI)
+---
 
-```ts
-import { estimatePrompt } from './loomlens-engine';
+## Features
 
-const result = estimatePrompt(
-  'Write a function that parses JSON and handles errors gracefully',
-  'minimax/MiniMax-M2'
-);
+- **19+ AI models** across 3 cost tiers
+- **Real-time token estimation** — input + cached + output
+- **Live cost comparison** — see exactly what each model costs before you run
+- **LoomToken balance** — connect your API key to see credit balance and charges
+- **Savings calculator** — shows how much you save vs. the most expensive option
 
-console.log(result.estimates[0].estimatedCost);  // $0.0012
-console.log(result.recommended?.modelName);       // 'Claude Sonnet 4'
-console.log(result.tip);                          // 'Using MiniMax M2...'
-```
+---
 
-## Architecture
+## How It Works
 
-```
-┌─────────────────────────────────────────────┐
-│         LoomLensEngine (engine.ts)          │
-│  ┌──────────────┐  ┌────────────────────┐  │
-│  │ TokenEstimator│  │ TaskTypeDetector   │  │
-│  │ (char heuristics)│ │ (keyword analysis │  │
-│  └──────────────┘  └────────────────────┘  │
-│  ┌──────────────┐  ┌────────────────────┐  │
-│  │ CostCalculator│  │ FitScorer          │  │
-│  │ ($/token × tp │  │ (capability match) │  │
-│  └──────────────┘  └────────────────────┘  │
-└────────────────────┬──────────────────────┘
-                     │ LoomLensResult
-        ┌────────────┼────────────┐
-        ▼            ▼            ▼
-  LoomLensPanel   /loomlens   POST /loomlens/override
-  (React UI)     (cmd)        → before_model_resolve hook
-```
+1. Paste your prompt
+2. Select models to compare
+3. Click **Estimate Cost**
+4. See exact cost per run for every model
 
-## Integration with Existing LoomLensAI-Live
+---
 
-The existing `signal-loom-live/panel.html` is a **post-hoc tracker** (shows cumulative cost after prompts fire).  
-LoomLens Live (`signal-loom-loomlens/`) is a **pre-flight estimator** (shows alternatives before you send).
+## LoomTokens
 
-They are complementary and can coexist. The OpenClaw plugin bridges them.
+Connect your API key to use pay-per-call estimates:
 
-## Next Steps
+- **100 LoomTokens** per estimate call = **$0.10**
+- Purchase credit packs: $5 (5,000) · $10 (10,000) · $25 (25,000) · $50 (50,000) · $100 (100,000)
+- Balance updates automatically after each call
 
-1. **Package** — add `package.json` + build step (tsc → JS bundle)
-2. **Test** — add unit tests for token estimation accuracy
-3. **Install plugin** — copy to `~/.openclaw/plugins/loomlens/`
-4. **Panel injection** — mount `loomlens-panel.tsx` into OpenClaw webchat DOM
-5. **Live pricing** — wire OpenRouter API for real-time pricing instead of static registry
+---
 
+Built by [Signal Loom AI](https://signalloomai.com) · [API Docs](https://signalloomai.com/docs)