tracecli 0.1.0__tar.gz

tracecli-0.1.0/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 Haseeb Arshad
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

tracecli-0.1.0/PKG-INFO

@@ -0,0 +1,136 @@
+Metadata-Version: 2.4
+Name: tracecli
+Version: 0.1.0
+Summary: The terminal's black box for your digital life. A privacy-first activity monitor.
+Author-email: Haseeb Arshad <haseebarshad@users.noreply.github.com>
+License: MIT
+Project-URL: Homepage, https://github.com/Haseeb-Arshad/trace-cli
+Project-URL: Repository, https://github.com/Haseeb-Arshad/trace-cli
+Project-URL: Issues, https://github.com/Haseeb-Arshad/trace-cli/issues
+Keywords: productivity,tracker,cli,windows,privacy,copilot
+Classifier: Development Status :: 4 - Beta
+Classifier: Environment :: Console
+Classifier: Operating System :: Microsoft :: Windows
+Classifier: Programming Language :: Python :: 3
+Classifier: Topic :: Utilities
+Requires-Python: >=3.9
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: pywin32>=306
+Requires-Dist: rich>=13.0
+Requires-Dist: click>=8.1
+Requires-Dist: psutil>=5.9
+Provides-Extra: dev
+Requires-Dist: pytest>=7.0; extra == "dev"
+Requires-Dist: pytest-cov; extra == "dev"
+Dynamic: license-file
+
+# TraceCLI
+
+**The terminal's black box for your digital life.**
+
+TraceCLI is a privacy-first, AI-powered activity tracker and productivity coach that lives in your terminal. It monitors your digital habits, actively protects your focus, and provides AI-powered coaching to improve your workflow—all without your data ever leaving your machine.
+
+[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](LICENSE)
+[![GitHub Repository](https://img.shields.io/badge/GitHub-trace--cli-blue.svg?logo=github)](https://github.com/Haseeb-Arshad/trace-cli)
+
+## Why TraceCLI?
+
+TraceCLI isn't just a time tracker. It's an intelligent agent designed to understand your workflow.
+
+- **Privacy First**: All data is stored locally in `~/.tracecli/trace.db`. Your activity data stays on your machine.
+- **Zero Friction**: Auto-starts with Windows. Tracks active windows, browser searches, and system resources automatically in the background.
+- **AI-Powered**: Chat with your productivity data using Natural Language. Get personalized insights and weekly digests to understand where your time goes.
+
+---
+
+## Key Features
+
+### Productivity Heatmap
+Visualize your consistency with a GitHub-style contribution graph.
+```bash
+tracecli heatmap --weeks 52
+```
+
+### Focus Mode
+A Pomodoro timer with active distraction detection. If you switch to a non-productive app (e.g., social media) during a session, TraceCLI alerts you immediately to get you back on track.
+```bash
+tracecli focus --duration 25 --goal "Core Feature Implementation"
+```
+
+### AI Insights
+Get personalized productivity digests or ask questions about your habits using natural language.
+```bash
+tracecli insights  # Weekly coaching report
+tracecli ask "What was my most used app on Monday?"
+```
+
+### Configurable Rules
+Define exactly what counts as "Productive" or "Distraction" for you using simple configuration rules.
+
+---
+
+## Installation
+
+You can install TraceCLI directly from PyPI (coming soon) or from source.
+
+```bash
+# Clone the repository
+git clone https://github.com/Haseeb-Arshad/trace-cli.git
+cd trace-cli
+
+# Install in editable mode
+pip install -e .
+
+# Verify installation
+tracecli --version
+```
+
+### AI Configuration (Optional)
+Supported providers: `gemini`, `openai`, `claude`.
+```bash
+tracecli config --provider gemini --key YOUR_API_KEY
+```
+
+---
+
+## Command Reference
+
+### Core & Tracking
+- `tracecli start`: Start activity tracking (use `--background` to run silently).
+- `tracecli live`: View real-time activity feed.
+- `tracecli status`: Check CLI and Database status.
+- `tracecli autostart enable`: Enable start on Windows login.
+
+### Analytics & Reports
+- `tracecli stats`: View daily productivity summary.
+- `tracecli heatmap`: GitHub-style productivity graph.
+- `tracecli report`: Detailed daily report with app breakdowns.
+- `tracecli timeline`: Visual daily timeline of activities.
+- `tracecli app [NAME]`: Deep dive into a specific application's usage.
+
+### System & Browsing
+- `tracecli urls`: Browser history and domain breakdown.
+- `tracecli searches`: Recent browser search queries.
+- `tracecli system`: System resource overview (CPU/RAM snapshots).
+
+---
+
+## Architecture
+
+TraceCLI is built on a modular architecture designed for performance and extensibility.
+
+- **Tracker**: Polls the Windows API to detect active applications and measures engagement.
+- **Categorizer**: Maps process names and window titles to categories (Development, Communication, Productivity, etc.) using a flexible rule engine.
+- **FocusMonitor**: A dedicated background thread that enforces focus rules during active sessions.
+- **BrowserBridge**: Safely reads browser history databases (Chrome/Edge) to provide context on web usage without invasive extensions.
+
+---
+
+## Contributing
+
+We welcome contributions! Please see `CONTRIBUTING.md` for details on how to get started.
+
+## License
+
+TraceCLI is released under the [MIT License](LICENSE).

tracecli-0.1.0/README.md

@@ -0,0 +1,109 @@
+# TraceCLI
+
+**The terminal's black box for your digital life.**
+
+TraceCLI is a privacy-first, AI-powered activity tracker and productivity coach that lives in your terminal. It monitors your digital habits, actively protects your focus, and provides AI-powered coaching to improve your workflow—all without your data ever leaving your machine.
+
+[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](LICENSE)
+[![GitHub Repository](https://img.shields.io/badge/GitHub-trace--cli-blue.svg?logo=github)](https://github.com/Haseeb-Arshad/trace-cli)
+
+## Why TraceCLI?
+
+TraceCLI isn't just a time tracker. It's an intelligent agent designed to understand your workflow.
+
+- **Privacy First**: All data is stored locally in `~/.tracecli/trace.db`. Your activity data stays on your machine.
+- **Zero Friction**: Auto-starts with Windows. Tracks active windows, browser searches, and system resources automatically in the background.
+- **AI-Powered**: Chat with your productivity data using Natural Language. Get personalized insights and weekly digests to understand where your time goes.
+
+---
+
+## Key Features
+
+### Productivity Heatmap
+Visualize your consistency with a GitHub-style contribution graph.
+```bash
+tracecli heatmap --weeks 52
+```
+
+### Focus Mode
+A Pomodoro timer with active distraction detection. If you switch to a non-productive app (e.g., social media) during a session, TraceCLI alerts you immediately to get you back on track.
+```bash
+tracecli focus --duration 25 --goal "Core Feature Implementation"
+```
+
+### AI Insights
+Get personalized productivity digests or ask questions about your habits using natural language.
+```bash
+tracecli insights  # Weekly coaching report
+tracecli ask "What was my most used app on Monday?"
+```
+
+### Configurable Rules
+Define exactly what counts as "Productive" or "Distraction" for you using simple configuration rules.
+
+---
+
+## Installation
+
+You can install TraceCLI directly from PyPI (coming soon) or from source.
+
+```bash
+# Clone the repository
+git clone https://github.com/Haseeb-Arshad/trace-cli.git
+cd trace-cli
+
+# Install in editable mode
+pip install -e .
+
+# Verify installation
+tracecli --version
+```
+
+### AI Configuration (Optional)
+Supported providers: `gemini`, `openai`, `claude`.
+```bash
+tracecli config --provider gemini --key YOUR_API_KEY
+```
+
+---
+
+## Command Reference
+
+### Core & Tracking
+- `tracecli start`: Start activity tracking (use `--background` to run silently).
+- `tracecli live`: View real-time activity feed.
+- `tracecli status`: Check CLI and Database status.
+- `tracecli autostart enable`: Enable start on Windows login.
+
+### Analytics & Reports
+- `tracecli stats`: View daily productivity summary.
+- `tracecli heatmap`: GitHub-style productivity graph.
+- `tracecli report`: Detailed daily report with app breakdowns.
+- `tracecli timeline`: Visual daily timeline of activities.
+- `tracecli app [NAME]`: Deep dive into a specific application's usage.
+
+### System & Browsing
+- `tracecli urls`: Browser history and domain breakdown.
+- `tracecli searches`: Recent browser search queries.
+- `tracecli system`: System resource overview (CPU/RAM snapshots).
+
+---
+
+## Architecture
+
+TraceCLI is built on a modular architecture designed for performance and extensibility.
+
+- **Tracker**: Polls the Windows API to detect active applications and measures engagement.
+- **Categorizer**: Maps process names and window titles to categories (Development, Communication, Productivity, etc.) using a flexible rule engine.
+- **FocusMonitor**: A dedicated background thread that enforces focus rules during active sessions.
+- **BrowserBridge**: Safely reads browser history databases (Chrome/Edge) to provide context on web usage without invasive extensions.
+
+---
+
+## Contributing
+
+We welcome contributions! Please see `CONTRIBUTING.md` for details on how to get started.
+
+## License
+
+TraceCLI is released under the [MIT License](LICENSE).

tracecli-0.1.0/pyproject.toml

@@ -0,0 +1,41 @@
+[build-system]
+requires = ["setuptools>=68.0", "wheel"]
+build-backend = "setuptools.build_meta"
+
+[project]
+name = "tracecli"
+version = "0.1.0"
+description = "The terminal's black box for your digital life. A privacy-first activity monitor."
+readme = "README.md"
+requires-python = ">=3.9"
+license = {text = "MIT"}
+authors = [{name = "Haseeb Arshad", email = "haseebarshad@users.noreply.github.com"}]
+keywords = ["productivity", "tracker", "cli", "windows", "privacy", "copilot"]
+
+classifiers = [
+    "Development Status :: 4 - Beta",
+    "Environment :: Console",
+    "Operating System :: Microsoft :: Windows",
+    "Programming Language :: Python :: 3",
+    "Topic :: Utilities",
+]
+dependencies = [
+    "pywin32>=306",
+    "rich>=13.0",
+    "click>=8.1",
+    "psutil>=5.9",
+]
+
+[project.urls]
+Homepage = "https://github.com/Haseeb-Arshad/trace-cli"
+Repository = "https://github.com/Haseeb-Arshad/trace-cli"
+Issues = "https://github.com/Haseeb-Arshad/trace-cli/issues"
+
+[project.optional-dependencies]
+dev = ["pytest>=7.0", "pytest-cov"]
+
+[project.scripts]
+tracecli = "src.cli:main"
+
+[tool.setuptools.packages.find]
+include = ["src*"]

tracecli-0.1.0/setup.cfg

@@ -0,0 +1,4 @@
+[egg_info]
+tag_build = 
+tag_date = 0
+

tracecli-0.1.0/src/__init__.py

@@ -0,0 +1,3 @@
+"""TraceCLI — The terminal's black box for your digital life."""
+
+__version__ = "2.0.0"

tracecli-0.1.0/src/__main__.py

@@ -0,0 +1,6 @@
+"""TraceCLI — Run as module: python -m src"""
+
+from .cli import main
+
+if __name__ == "__main__":
+    main()

tracecli-0.1.0/src/ai.py

@@ -0,0 +1,368 @@
+"""
+TraceCLI AI Agent
+~~~~~~~~~~~~~~~~~
+Zero-dependency AI integration using urllib.
+Handles Text-to-SQL generation and natural language summarization.
+"""
+
+import json
+import sqlite3
+import urllib.request
+import urllib.error
+from datetime import date
+from typing import Optional, Dict, Any
+
+from rich.console import Console
+from rich.panel import Panel
+
+from . import config
+from .database import get_connection, DB_PATH
+
+console = Console()
+
+# ── API Clients ────────────────────────────────────────────────────────────
+
+def _post_json(url: str, headers: Dict[str, str], data: Dict[str, Any]) -> Dict:
+    """Helper to send JSON POST request using standard library (with retries)."""
+    import time
+    
+    max_retries = 3
+    backoff = 1  # seconds
+
+    for attempt in range(max_retries):
+        try:
+            req = urllib.request.Request(
+                url,
+                data=json.dumps(data).encode("utf-8"),
+                headers=headers,
+                method="POST"
+            )
+            with urllib.request.urlopen(req) as response:
+                return json.load(response)
+        except urllib.error.HTTPError as e:
+            if e.code == 429:
+                if attempt < max_retries - 1:
+                    console.print(f"[yellow]Rate limited (429). Retrying in {backoff}s...[/yellow]")
+                    time.sleep(backoff)
+                    backoff *= 2
+                    continue
+            
+            try:
+                err_body = e.read().decode()
+                console.print(f"[red]API Error {e.code}: {err_body}[/red]")
+            except Exception:
+                console.print(f"[red]API Error: {e}[/red]")
+            return {}
+        except Exception as e:
+            console.print(f"[red]Network Error: {e}[/red]")
+            return {}
+    return {}
+
+
+def call_gemini(api_key: str, prompt: str, model: str = ""):
+    """Call Google Gemini API."""
+    model = model or "gemini-flash-latest"
+    url = f"https://generativelanguage.googleapis.com/v1beta/models/{model}:generateContent?key={api_key}"
+    headers = {"Content-Type": "application/json"}
+    data = {
+        "contents": [{"parts": [{"text": prompt}]}]
+    }
+    
+    resp = _post_json(url, headers, data)
+    try:
+        return resp["candidates"][0]["content"]["parts"][0]["text"]
+    except (KeyError, IndexError):
+        return None
+
+
+def call_openai(api_key: str, prompt: str, model: str = ""):
+    """Call OpenAI API."""
+    model = model or "gpt-3.5-turbo"
+    url = "https://api.openai.com/v1/chat/completions"
+    headers = {
+        "Content-Type": "application/json",
+        "Authorization": f"Bearer {api_key}"
+    }
+    data = {
+        "model": model,
+        "messages": [{"role": "user", "content": prompt}],
+        "temperature": 0
+    }
+    
+    resp = _post_json(url, headers, data)
+    try:
+        return resp["choices"][0]["message"]["content"]
+    except (KeyError, IndexError):
+        return None
+
+
+def call_claude(api_key: str, prompt: str, model: str = ""):
+    """Call Anthropic Claude API."""
+    model = model or "claude-3-haiku-20240307"
+    url = "https://api.anthropic.com/v1/messages"
+    headers = {
+        "x-api-key": api_key,
+        "anthropic-version": "2023-06-01",
+        "content-type": "application/json"
+    }
+    data = {
+        "model": model,
+        "max_tokens": 1024,
+        "messages": [{"role": "user", "content": prompt}]
+    }
+    
+    resp = _post_json(url, headers, data)
+    try:
+        return resp["content"][0]["text"]
+    except (KeyError, IndexError):
+        return None
+
+
+def ask_llm(provider: str, api_key: str, prompt: str, model: str = "") -> Optional[str]:
+    """Dispatch to the configured provider."""
+    provider = provider.lower()
+    if provider == "gemini":
+        return call_gemini(api_key, prompt, model)
+    elif provider == "openai":
+        return call_openai(api_key, prompt, model)
+    elif provider == "claude":
+        return call_claude(api_key, prompt, model)
+    else:
+        console.print(f"[red]Unknown provider: {provider}[/red]")
+        return None
+
+
+# ── Text-to-SQL Logic ──────────────────────────────────────────────────────
+
+def get_schema_summary() -> str:
+    """Get a summary of the database schema for the LLM."""
+    conn = get_connection()
+    cursor = conn.cursor()
+    
+    schema = []
+    tables = ["activity_log", "daily_stats", "process_snapshots", "browser_urls", "search_history"]
+    
+    for table in tables:
+        cursor.execute(f"PRAGMA table_info({table})")
+        cols = [f"{row['name']} ({row['type']})" for row in cursor.fetchall()]
+        schema.append(f"Table {table}: " + ", ".join(cols))
+        
+    return "\n".join(schema)
+
+
+VALID_CATEGORIES = [
+    "💻 Development", "🌐 Browsing", "📚 Research", "💬 Communication",
+    "📝 Productivity", "🎮 Distraction", "❓ Other"
+]
+
+
+def generate_sql(question: str, provider: str, api_key: str, model: str) -> Optional[str]:
+    """Ask LLM to generate SQL."""
+    schema = get_schema_summary()
+    today = date.today().isoformat()
+    
+    prompt = f"""
+    You are a SQLite expert. Given the database schema below, write a single SQL query to answer the user's question.
+    
+    Current Date: {today}
+    Schema:
+    {schema}
+    
+    Rules:
+    1. Return ONLY the SQL query. No markdown, no explanations.
+    2. Use safe localized time comparison: `start_time LIKE '2023-01-01%'` or `timestamp LIKE '2023-01-01%'`.
+    3. Calculate durations in minutes if relevant (duration_seconds / 60 or visit_duration / 60).
+    4. Only use SELECT statements.
+    
+    Data Quirks:
+    - `category` column uses these exact strings: {VALID_CATEGORIES}
+    - If user asks for "Entertainment", "Social", or "Chat", look for keywords in `app_name` OR include `category = '❓ Other'` to find uncategorized apps.
+    - `search_history` entries with `url = ''` are captured from browser titles.
+    - To join `search_history` with `browser_urls` for productivity time, use:
+      `ON s.url = b.url OR (s.url = '' AND b.title LIKE '%' || s.query || '%')`
+    
+    Question: {question}
+    SQL:
+    """
+    
+    response = ask_llm(provider, api_key, prompt, model)
+    if not response:
+        return None
+        
+    # Clean up response (remove markdown if any)
+    sql = response.strip()
+    if sql.startswith("```"):
+        sql = sql.split("\n", 1)[1]
+    if sql.endswith("```"):
+        sql = sql.rsplit("\n", 1)[0]
+    return sql.strip()
+
+
+def summarize_result(question: str, sql: str, rows: list, provider: str, api_key: str, model: str) -> str:
+    """Ask LLM to summarize the SQL result."""
+    # Truncate rows if too large
+    data_str = str(rows[:20])
+    if len(rows) > 20:
+        data_str += f"... (+{len(rows)-20} more rows)"
+        
+    prompt = f"""
+    You are a helpful assistant analyzing activity data.
+    User Question: "{question}"
+    SQL Query Run: "{sql}"
+    Result Data: {data_str}
+    
+    You are a helpful assistant analyzing activity data.
+    User Question: "{question}"
+    SQL Query Run: "{sql}"
+    Result Data: {data_str}
+    
+    Provide a concise, friendly answer based on the data.
+    - If you see apps with `category='❓ Other'`, analyze their `app_name` or `window_title`. If they look like games, social media, or entertainment, point them out as potential distractions.
+    - If the data is empty, say so politely.
+    Answer:
+    """
+    
+    return ask_llm(provider, api_key, prompt, model) or "Error generating summary."
+
+
+def handle_ask(question: str):
+    """Main entry point for 'tracecli ask'."""
+    provider, api_key, model = config.get_ai_config()
+    
+    if not api_key:
+        console.print("[yellow]⚠️  AI API Key not configured.[/yellow]")
+        console.print(f"Run [bold]tracecli config --key YOUR_KEY[/bold] to set it up.")
+        return
+
+    console.print(f"[dim]Thinking with {provider}...[/dim]")
+    
+    # 1. Generate SQL
+    sql = generate_sql(question, provider, api_key, model)
+    if not sql:
+        console.print("[red]Failed to generate SQL query.[/red]")
+        return
+        
+    # Safety Check
+    if not sql.upper().startswith("SELECT"):
+        console.print("[red]Safety Violation: Generated SQL is not a SELECT statement.[/red]")
+        return
+        
+    console.print(f"[dim]Executing: {sql}[/dim]")
+    
+    # 2. Execute SQL
+    try:
+        conn = get_connection()
+        # Ensure read-only by using a cursor? sqlite3 doesn't easily enforce read-only per query
+        # But we checked for SELECT.
+        cursor = conn.execute(sql)
+        cols = [description[0] for description in cursor.description]
+        rows = [dict(zip(cols, row)) for row in cursor.fetchall()]
+    except Exception as e:
+        console.print(f"[red]SQL Error: {e}[/red]")
+        return
+
+    if not rows:
+        console.print("\n[yellow]No data found matching your query.[/yellow]\n")
+        return
+
+    # 3. Summarize
+    console.print("[dim]Analyzing results...[/dim]")
+    answer = summarize_result(question, sql, rows, provider, api_key, model)
+    
+    console.print()
+    console.print(Panel(answer, title="🤖 AI Answer", border_style="cyan"))
+    console.print()
+
+
+# ── Weekly Productivity Digest ─────────────────────────────────────────────
+
+def generate_weekly_digest(days: int = 7) -> Optional[str]:
+    """
+    Generate an AI-powered productivity digest for the past N days.
+
+    Collects daily stats, app usage, and search patterns, then asks
+    the LLM to generate a personalized productivity coaching report.
+    """
+    provider, api_key, model = config.get_ai_config()
+    if not api_key:
+        return None
+
+    from .database import (
+        get_stats_range, get_category_breakdown, get_app_breakdown,
+        get_focus_stats, query_searches,
+    )
+    from datetime import date, timedelta
+
+    # Gather data
+    stats = get_stats_range(days)
+    today = date.today()
+
+    if not stats:
+        return "No activity data found for the past week."
+
+    # Build structured summary
+    total_seconds = sum(s["total_seconds"] for s in stats)
+    productive_seconds = sum(s["productive_seconds"] for s in stats)
+    distraction_seconds = sum(s["distraction_seconds"] for s in stats)
+
+    # App breakdown for latest day
+    app_breakdown = get_app_breakdown(today)
+    top_apps = [
+        {"app": a["app_name"], "time_hours": round(a["total_seconds"] / 3600, 1)}
+        for a in app_breakdown[:5]
+    ]
+
+    # Category breakdown for latest day
+    cat_breakdown = get_category_breakdown(today)
+    categories = [
+        {"category": c["category"], "time_hours": round(c["total_seconds"] / 3600, 1)}
+        for c in cat_breakdown
+    ]
+
+    # Focus stats
+    focus = get_focus_stats()
+
+    # Search patterns
+    searches = query_searches(today)
+    recent_queries = [s["query"] for s in searches[:10]]
+
+    summary = {
+        "period_days": days,
+        "total_tracked_hours": round(total_seconds / 3600, 1),
+        "productive_hours": round(productive_seconds / 3600, 1),
+        "distraction_hours": round(distraction_seconds / 3600, 1),
+        "productivity_score": round((productive_seconds / total_seconds) * 100) if total_seconds > 0 else 0,
+        "days_tracked": len(stats),
+        "daily_breakdown": [
+            {
+                "date": s["date"],
+                "hours": round(s["total_seconds"] / 3600, 1),
+                "productive_hours": round(s["productive_seconds"] / 3600, 1),
+                "score": round((s["productive_seconds"] / s["total_seconds"]) * 100) if s["total_seconds"] > 0 else 0,
+            }
+            for s in stats
+        ],
+        "top_apps": top_apps,
+        "categories": categories,
+        "focus_sessions": focus.get("total_sessions", 0),
+        "avg_focus_score": round(focus.get("avg_focus_score", 0), 1),
+        "recent_searches": recent_queries,
+    }
+
+    prompt = f"""You are a friendly, insightful productivity coach analyzing someone's computer usage data.
+
+Here is their activity summary for the past {days} days:
+{json.dumps(summary, indent=2)}
+
+Generate a concise, personalized productivity digest with these sections:
+1. 🏆 **Top Achievement** — Highlight their best metric or pattern (be specific with numbers)
+2. ⚠️ **Biggest Distraction Pattern** — Identify when/where they get distracted most
+3. 💡 **Actionable Suggestion** — One specific, practical tip to improve tomorrow
+4. 📊 **Week-over-Week** — Comment on trends if multiple days of data exist
+
+Keep each section to 1-2 sentences. Be encouraging but honest.
+Use emoji and make it engaging. Don't use markdown headers, just the emoji labels above.
+If there's limited data, acknowledge it and focus on what you can see."""
+
+    result = ask_llm(provider, api_key, prompt, model)
+    return result