# AmoxSQL v1.8.0 — Data Visualization Overhaul 📊

[DSandovalFlavio]

The Modern Codex for Local Data Analysis

🌐 Official Website · 💖 Sponsor this project

---

What is AmoxSQL?

A professional, local-first IDE for DuckDB. Built for data analysts and engineers who need speed, privacy, and powerful tooling — no cloud required.

---

✨ What's New in v1.8.0

📈 Massive Chart System Rewrite

We completely tore down the old monolithic visualization engine (a massive 2,200+ line component) and rebuilt it from the ground up into a highly modular, performant architecture. The new DataVisualizer is now split into 15+ specialized modules, powered by a robust useReducer state engine for lightning-fast UI updates.

📊 5 New Chart Types

Alongside the existing Columns, Lines, Scatters, and Donuts, v1.8.0 introduces five powerful new ways to visualize your Data:

• Combo Charts (Bar + Line): The most requested feature! Combine bars and lines on the same chart with dual Y-axis support.
• Heatmaps: Perfect for cohort analysis and retention tables, mapping color intensity directly to data values.
• Funnel Charts: Visualize conversion rates across sequential pipeline stages.
• Bubble Charts: Add a third dimension to your scatter plots by mapping a metric to bubble size.
• Stacked Areas: True proportional stacked area charts for time-series part-to-whole analysis.

🎨 Advanced Premium Customization

We've brought BI-level styling directly into your local IDE:

• Visual Chart Selector: A beautiful new icon-grid makes choosing the right chart type intuitive.
• Color & Typography: 18 curated color palettes, custom background tones, and 7 professional fonts (Inter, Roboto, JetBrains, etc.) with a global text scale slider.
• Smart Data Labels: Labels now feature automatic contrast detection (white text on dark bars, dark text on light bars) and an intelligent "auto-hide" slider that prevents overlapping when bars get too small.
• Headline KPIs: Add a massive, prominent Metric Number above your chart that automatically calculates Deltas (▲/▼) against previous periods.
• Storytelling Annotations: Add customizable Goal Lines, Linear Trends, and Moving Averages with just a click.

🛠️ UX & Stability Improvements

• Tab Reorganization: The visualization config panels now follow a logical flow: 📊 Chart → 📋 Data → ⚙ Detail → 📐 Axes → 🎨 Theme → ✏️ Story.
• Backward Compatibility: All your existing .amoxvis chart configurations from older versions will load seamlessly into the new engine thanks to a transparent migration layer.
• Performance: npm run dist now builds the client 2x faster with 0 errors.

---

🧠 Full Feature Set

Everything from previous releases, plus the new additions above:

• 🏗️ DBT Studio — Full Module for Data Engineers with smart Conda Environment detection, Model Generators, and local compilation.
• 🤖 AI-Powered SQL — Local Mode (Ollama, 100% offline) + Cloud Mode (Google Gemini). Natural Language to SQL with full schema awareness.
• 📝 SQL Notebooks — Jupyter-like .sqlnb files with Markdown + SQL cells. Presentation Mode, PDF Export, and interactive HTML Export.
• 📊 Advanced Visualization — Combo, Heatmap, Funnel, Bubble, Line, Bar, Scatter, Donut charts with persistent .amoxvis configs.
• 🐛 Debugging Tools — CTE Step-Through Debugger and Query Execution Plan viewer with bottleneck detection.
• 💾 Database Management — In-Memory or persistent DuckDB connections. BigQuery-style table inspector.
• 📥 Data IO — CSV/Parquet/Excel bulk import. Export results as CSV, JSON, Parquet, or Excel via DuckDB's native COPY TO.
• 🧩 Extension Gallery — Browse, search, and install DuckDB extensions visually.

---

📥 Download

Download the installer or build from source:

👇 Download v1.8.0 from Releases

---

🔮 What's Next?

Future versions will continue expanding the DBT Studio module, adding more data engineering workflows, and refining the overall IDE experience. Pre-built installers are available exclusively to GitHub Sponsors. The source code will always remain public.

---

Built with 💙 from Latin America to the World.

---

This discussion was created from the release # AmoxSQL v1.8.0 — Data Visualization Overhaul 📊.

[mergisi]

The dual-mode AI approach (Ollama local + Gemini cloud) for the NL-to-SQL layer is a smart architecture decision. Giving users the privacy-first option with local inference while keeping cloud as a fallback covers both enterprise and individual use cases cleanly.

The CTE step-through debugger caught my attention — that's a feature most SQL IDEs still lack. CTEs are where query complexity tends to hide, especially when people chain 4-5 of them together. Being able to inspect intermediate results at each step changes how you debug recursive queries and window function pipelines.

One thing I'm curious about: how does the schema awareness in the AI-Powered SQL feature handle DuckDB-specific syntax like QUALIFY, EXCLUDE, or COLUMNS(*) expressions? These are powerful but uncommon enough that most NL-to-SQL models under-represent them in training data. If you're fine-tuning or using structured prompts with the schema context, that could be a differentiator worth highlighting.

For those exploring the NL-to-SQL space more broadly, ai2sql.io is another useful tool — it converts natural language to SQL and supports multiple dialects, which helps when learning syntax differences across engines.