---

Add to Whittl itself. A plugin is a folder with a small manifest and a Python file. It can add a tool the AI calls while it works, a tab beside Code, a status bar readout, a dialog or a menu item under Edit > Plugins, a skill, or code that runs when a generation starts or ends, a project opens, or you run your app. Whittl's own Assets, Archives, Terminal and Help tabs are plugins built the same way.

• Whittl writes them for you: File > New Project, pick Plugin, and describe what you want
• Checked before anything runs: a tool's examples are run, and a panel is built first in an invisible copy of Qt, so a broken plugin is held back instead of crashing the window
• Plugins you add stay off until you switch them on, one by one, in Edit > Plugins
• Plugins you write are yours to share or sell under any license, and the templates and bundled plugins you start from are MIT No Attribution, so nothing from them needs crediting
• A full example plugin ships as a template, and the docs walk from a two-file hello world to your own tab

---

Every backend gets the same tools, and what a model is given depends on what it can do, not on who makes it. A cheap model, a local one and a frontier one all work on your project the same way.

• A model that calls tools works through your project step by step: it reads what it needs, makes targeted edits, runs your app and your tests, and fixes what it broke. Files stream into the editor as they are written
• A model that cannot call tools reliably writes short scripts instead, which make the same edits through a small, safe API, so smaller and local models can work on a project too
• The AI can look at a desktop app it built: it opens it invisibly, reads the window, clicks buttons, and can show you a picture of it in the chat
• Stop stops. A model stuck repeating itself is ended instead of watched, and a turn that keeps reading without learning anything new is stopped
• An overloaded or failing provider is retried and named as the cause, and a turn that fails partway keeps the edits it made instead of regenerating over them

---

Every other AI coding tool is a chat window over someone else's model. Whittl is the knowledge layer that sits between you and the AI — 75+ auto-fix rules, a curated skills library for PySide6, Flet, CustomTkinter, and general Python, oscillation guards, hard round caps, tool executors, and validators. The model writes the code; Whittl makes the model better at building Python desktop apps.

That layer compounds with use. Every auto-fix rule learned from a real bug. Every skill written from a confirmed pattern. The AI model lineup changes monthly — Sonnet, GPT-5, DeepSeek V4, Qwen3.5 Plus — the knowledge layer keeps Whittl's output consistent and reliable across all of them.

• 75+ auto-fix rules covering Qt scoped enums, Flet hallucinations, PyInstaller resource paths, deprecated APIs, hallucinated class names, and dozens more — Whittl catches and rewrites what the AI gets wrong before the code ever fails
• Skill library auto-injected based on your project's framework — PySide6 patterns for desktop apps, Flet mobile rules for Android builds, threading rules for both
• Oscillation guard + 5-round cap prevent runaway autofix loops that would otherwise burn tokens forever on the same error
• Validators catch bad code before it reaches your project — duplicate marker handling, file truncation detection, structural diff guards
• Built-in skills for QPainter drawing, named visual styles, games, tables, files and drag and drop, settings that survive a restart, serial ports and 3D with Qt3D, every rule checked on the real libraries before it was written
• When the AI catches and fixes its own mistake before finishing, the lesson goes into the auto-learned library, just as auto-fixes do
• Compatible with Claude Code skills: skill folders written for Claude Code are read the way Claude Code stores them, alongside Whittl's own library

---

Drop a screenshot of any UI into Whittl and get a native Python app. A Figma mock, a web app you want to run locally, your own music player you're tired of configuring. Every other AI coding tool outputs more web. Whittl uniquely converts design references into runnable software you own.

• Drop or paste one screenshot or several into a message
• Works with any vision-capable model on any backend, local Ollama vision models included. A model that cannot read images gets a description from one that can
• The pictures reach the AI on every turn of the conversation and every tool round, not just the first message, so it sees your reference while it builds
• Each picture shows as a thumbnail before you send, removable, and in the chat afterwards; click one to see it full size. Pictures survive a restart

---

• Code editor with syntax highlighting, AST-powered code folding, autocomplete, and bracket auto-pairing
• Real-time error detection (pyflakes) with inline error indicators
• Multi-file project management with tabbed editing and line count tracking
• Smart file routing — keyword search + AI selects the right files in large projects
• One-click builds — Windows .exe, Linux AppImage, Android APK, and setup installers
• Explain mode — highlight any code and get a plain-English explanation
• Undo/history — roll back bad generations to any previous state
• Multi-file search across your entire project
• In-app auto-updater — Download → Restart to install, with SHA-256 checksum verification, no browser hop
• Real .py files on your disk — open in any editor, version with git, own forever
• Edit project files in another editor, or with another AI tool, while Whittl is open: your changes and new files are kept, not overwritten
• Works on Windows 10 and 11, and on Linux (the AppImage runs on Ubuntu 22.04, Linux Mint 21, Debian 12 and newer)

---

• 5 targets: Desktop (PySide6), Mobile (Flet), Basic (CustomTkinter), General Python (CLI tools and scripts), and Whittl Plugin
• Conversational AI — ask questions, get explanations, iterate on your code through chat
• 5 AI backends — Claude, DeepSeek, Gemini, OpenRouter (200+ models, free ones included), or fully offline with Ollama
• Smart editing — the AI changes only the lines that need to change, dramatically faster and cheaper than rewriting whole files
• Self-healing autofix: when your app crashes, the AI reads the traceback, fixes the code and re-runs the app, up to three attempts after a generation and one when it crashes while you are using it
• Web research: the AI can search the web, read pages and save the files it needs while it works (Agent Mode)
• Test gate: if your project has tests, they run after the AI says it is done, and anything the change broke is reported before you see "done"
• Built-in cost cache — multi-round conversations reuse context so your API bill stays small
• Know what you are paying: a running cost readout, a gauge for how full the model's context window is, and an optional Usage tab that charts spending by day and model
• 75+ auto-fix rules catch common mistakes — hallucinated APIs, deprecated patterns, missing imports
• Context memory — the AI remembers your earlier conversation instead of forgetting after a few messages
• Bring your own API key — no middleman markup, no message limits, no monthly quotas

---

Mobile APK builds are currently experimental. They work but may need troubleshooting on some devices. Desktop builds (Windows/Linux) are stable and production-ready.

• Build Android APKs with one click — no Android Studio, no Java, no Gradle
• Device format selector — Phone, Tablet, or custom dimensions
• One-click install to device via USB
• Pure Python APK signing — all build tools downloaded automatically

---

---

Cormorant Garamond and Inter typography, Phosphor Duotone icons throughout, and a refined navy, cream, tan, and copper palette. Dark and light themes are both polished, and flip cleanly without stale colors bleeding through.

---

---

1. Start a new project or import existing Python code

2. Write code in the editor, or describe what you want and let AI generate it

3. Use the AI chat to ask questions, explain code, or iterate on changes

4. Run and test directly inside Whittl with built-in error detection

5. Build to .exe, AppImage, or Android APK with one click

---

---

---