Vibe coding the vibe coder - VibeSlicer

[Jody Harris]

Okay kiddos, here we go. 

Since I'm out of Claude tokens for the next two hours, I asked Grok Expert and Gemini Pro Extended the following question: 

Help me craft a prompt for Claude Fable 5. To create a full featured modern 3D printer g code slicer for fdm 3D printers with a graphical user interface, multi-threaded, support for STL, 3MF, STEP and OBJ files. Ability to import prusa slicer printer and filament profiles as well as orca slicer printer and filament profiles. Duplicate all prusa slicer and orca slicer features.

Grok's response: 

Here's a high-quality, ready-to-copy prompt optimized for Claude (Sonnet 3.5 / Opus or newer). It’s structured for iterative development, which is the only realistic way to build something this complex.Copy and paste this directly into Claude:markdown

You are a world-class software architect and developer with deep expertise in computational geometry, 3D printing, and large-scale desktop applications. You have previously worked on slicer engines and understand PrusaSlicer and OrcaSlicer internals very well.

Your task is to build **Fable 5**, a modern, full-featured, production-grade FDM G-code slicer from scratch.

### Hard Requirements

- **GUI**: Professional, modern interface using **PySide6** (Qt 6). Dark theme by default, clean layout, dockable panels, searchable settings, good tooltips, and responsive behavior.

- **Performance**: Fully multi-threaded. All slicing, mesh processing, and heavy calculations must run in background threads (`QThread` or `QThreadPool`) so the UI never freezes.

- **File Support**:

  - Import: STL (binary + ASCII), OBJ, 3MF, STEP

  - Export: G-code (.gcode)

- **Profile Import**:

  - Full ability to import **printer profiles** and **filament profiles** from both PrusaSlicer and OrcaSlicer.

  - Accurately parse and map their configuration parameters.

- **Feature Parity**: Replicate **all major and minor features** from the latest versions of PrusaSlicer and OrcaSlicer (as of 2025–2026). This includes (but is not limited to):

  - All infill patterns (rectilinear, grid, triangles, stars, cubic, gyroid, honeycomb, lightning, monotonic, etc.)

  - Tree/organic supports + classic supports + paint-on support capability

  - Variable layer height / adaptive layers

  - All seam options, ironing, fuzzy skin, scarf joints, monotonic infill, etc.

  - Full support settings (interface layers, support on build plate only, etc.)

  - Brim, skirt, raft with all parameters

  - Per-object settings and modifiers where possible

  - Advanced preview (feature type, speed, flow, layer-by-layer playback)

  - Accurate time and filament usage estimation

  - All speed, acceleration, retraction, cooling, and temperature settings

### Technology Stack (use this unless you have a clearly better proposal)

- Python 3.12+

- GUI: PySide6

- 3D visualization & mesh handling: **PyVista** (embedded in Qt)

- Geometry & polygon operations: `pyclipper`, `shapely`, `numpy`, `scipy`

- Optional but recommended: `trimesh`, `python-occ-core` (for STEP)

- Settings management: Pydantic models or well-structured dataclasses

- Threading: Qt’s signal/slot system + `concurrent.futures` where appropriate

### Development Philosophy

This is a large project. We will build it **iteratively** over multiple conversations. Never try to write the entire slicer in one response.

For this first message, do the following **in order**:

1. **Confirm or improve** the technology stack and high-level architecture.

2. **Design the architecture** — list the main modules/classes you will create (e.g. `MainWindow`, `SlicerEngine`, `SettingsManager`, `ProfileImporter`, `MeshProcessor`, `LayerSlicer`, `PathGenerator`, `GCodeWriter`, `PreviewManager`, etc.).

3. **Create the project structure** with a proper `pyproject.toml` / `requirements.txt`, `README.md`, and folder layout.

4. **Implement the foundation**:

   - Main application entry point

   - Main window with modern layout (menu bar, toolbar, central 3D view, dockable object list + settings panel)

   - PyVista 3D viewer embedded in Qt with basic model loading (STL + OBJ first)

   - Basic settings panel (tree or categorized form)

   - `SlicerEngine` class skeleton with a threaded `slice()` method and progress signals

   - Profile importer skeleton that can parse PrusaSlicer/OrcaSlicer `.ini` files (start with a few key parameters and make it extensible)

   - Simple G-code writer that outputs valid, commented G-code

5. Use **type hints**, clear docstrings, and clean modular code.

6. After delivering the initial code, tell me exactly what we should build next (in priority order) and ask which direction I want to go.

Make the code production-quality from the start — no “demo” shortcuts that will need to be rewritten later.

Begin now.

Why this prompt works well with Claude:It sets realistic expectations (iterative development).

It forces Claude to think architecturally first instead of dumping incomplete code.

It explicitly mentions feature parity with Prusa/Orca so Claude pulls from its knowledge of those programs.

It gives Claude a clear structure for the first response.

It leaves room for Claude to suggest better libraries if it has good ideas.

Optional Enhancements (you can add these if you want)You can append any of these to the prompt if desired:Add support for paint-on supports/seams early.

Make the settings system use the exact same parameter names/groups as OrcaSlicer where possible (easier profile importing).

Include a basic G-code simulator in the preview from the beginning.

Prioritize tree supports and gyroid infill early.

Would you like me to create a more aggressive/single-shot version of the prompt (for when you want Claude to go deeper in one response), or would you prefer any modifications to the current one (different tech stack, more/less emphasis on certain features, etc.)?

[Python?]

Gemini's response: 

Building a modern slicer from scratch is an incredibly ambitious and exciting project. To be completely candid, though, no current AI model—even an advanced or hypothetical future version like "Claude Fable 5"—can generate a fully featured, production-ready slicer with 100% PrusaSlicer and OrcaSlicer parity in a single prompt.

Modern slicers represent hundreds of thousands of lines of highly optimized C++ code. They rely heavily on complex computational geometry libraries (to handle non-manifold meshes, precise boolean operations, and routing), sophisticated path-planning algorithms, and extensive hardware testing.

However, an advanced LLM is the perfect tool to **architect the foundation**, set up the multi-threading pipeline, build the GUI scaffold, and draft the core geometry and profile-parsing modules. You will need to build this iteratively.

Here is a highly detailed, structured prompt designed to get the best possible architectural scaffolding and initial codebase out of an advanced AI.

### The Prompt to Copy & Paste

**System Prompt / Persona:**

Act as an expert C++ Software Architect, Computational Geometer, and UI/UX developer specializing in 3D graphics and CAM (Computer-Aided Manufacturing) software. Your task is to architect and scaffold a next-generation, high-performance 3D printer g-code slicer for FDM printers.

**Project Overview:**

I am building a modern, multi-threaded FDM slicer from scratch. It must feature a responsive Graphical User Interface, support standard mesh and CAD file formats natively, and achieve feature parity with PrusaSlicer and OrcaSlicer, including their complex toolpath generation and multi-material capabilities.

**Core Requirements:**

 1. **Tech Stack:** Propose a modern tech stack. Use C++20 or newer. Suggest the best libraries for the GUI (e.g., Qt or Dear ImGui), 3D rendering (Vulkan or OpenGL), and multi-threading (e.g., Intel TBB or standard threads).

 2. **File Parsing:** The engine must ingest STL, OBJ, 3MF, and native STEP files. Specify the C++ libraries required to parse these (e.g., OpenCASCADE for STEP, libigl or CGAL for meshes) and draft the unified internal data structure that will hold this geometry.

 3. **Profile Ingestion:** The software must natively import, parse, and translate PrusaSlicer and OrcaSlicer printer, filament, and print profiles (which are primarily JSON/INI-based).

 4. **Multi-Threading Pipeline:** The core slicing operations (layer generation, perimeter routing, infill generation, and support generation) must be fully multi-threaded to minimize processing time.

 5. **Advanced FDM Features:** Ensure the architecture natively supports complex multi-extruder setups, such as routing dissimilar polymers for dedicated interface layers, dynamic layer heights, and advanced pressure advance/linear advance calibrations found in Orca.

 6. **Cross-Platform Build:** Provide the initial CMakeLists.txt configured for cross-platform compilation, ensuring robust native builds for Linux, Windows, and macOS.

**Your Deliverables for Phase 1:**

 1. **High-Level Architecture Document:** Map out the core modules (UI, Geometry Engine, Slicing Engine, G-Code Generator, Profile Manager) and how data flows between them.

 2. **Library Stack:** A definitive list of the open-source C++ dependencies required to make this work.

 3. **Core Scaffolding:** Write the foundational C++ header files defining the primary classes for the Model, PrintProfile, Layer, and the SlicingEngine interface.

 4. **Profile Parser Draft:** Write a C++ function that demonstrates how to deserialize a standard OrcaSlicer filament JSON profile into our internal data structure.

Please prioritize code safety, scalability, and performance in your architectural design. Do not attempt to write the entire application at once; focus strictly on a robust, compiling Phase 1 scaffold.

### How to use this effectively:

 * **Iterative Development:** Once the AI provides the initial scaffolding and CMake files, verify that it compiles. Then, feed the AI one specific module at a time to build out (e.g., "Now, let's implement the OpenCASCADE wrapper to ingest a STEP file and convert it to our internal mesh format.").

 * **The Parity Challenge:** Duplicating *every* feature of Prusa and Orca is a massive undertaking. Focus on getting the core slicing loop (Import -> Slice -> Infill -> Gcode) working first, then use the AI to incrementally add complex features like organic trees, arachne perimeter generation, and scarf seams.

 * **Profile Translation:** Prusa and Orca use slightly different math and variable names for their configuration files. You will likely need to spend significant time having the AI write translation matrices to map their settings to your new slicer's internal logic.

Now, can we work together to craft the prompt to get a new, hopefully better, slicer? 

-j

Think carefully

[david merten]

What a mouth full. But very interesting reading. I have been thinking about an "engine lathe" style 3d printer that would rotate the part while the hot end travels along the length on a rail that raises up. I will need to make a special slicer for it. Something I wouldn't consider without the help of AI.    

               David

--
You received this message because you are subscribed to the Google Groups "3D Printing Tips and Tricks" group.
To unsubscribe from this group and stop receiving emails from it, send an email to 3d-printing-tips--...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/3d-printing-tips--tricks/SDJY8147gLAKmnzNlduIgcYDuLGH-10UDPv24CIV-DwBi3ljixlPT-YG9tn35p4xXGHmifw-nPX0SqtccnZZ-t6BDfh13bZDDzuCUr1piII%3D%40harrisdev.com.

[vr...@optonline.net]

OMG - Jody - BRING IT ON!!!!

The Challenge is Viable and Just - Bring it on - Let's use AI to build a Superior Slicer!

We need to give it MORE Functions than current slicers!

 

But - I have a Challenge that would make a new slicer like this - to be Beyond what has already been done!

 

Why Do I say this???

 

Cause - I DO know a 5-axis type of printer - that can swivel and Dance the build plate - kinda like that other NEW printer that was posted in this forum. But - the one I speak of - is NOT NEW - has been around for a WHILE - and it would be INSANELY Amazing to make it do 5D kinda prints - but, IN HIGH TEMP Materials like Ultem and PEEK!!!!

 

Ready...

Set...

Go!!!!!!!!!!!!!!!!!!

 

-K

 

-- 

You received this message because you are subscribed to the Google Groups "3D Printing Tips and Tricks" group.

To unsubscribe from this group and stop receiving emails from it, send an email to unsub...@googlegroups.com">3d-printing-tips--tricks+unsub...@googlegroups.com.

[Jody Harris]

When my 5 hours were up, I fed the prompt prompt to Claude Opus 4.8 Max. I thought Claude might have a stroke. While it chewed, I read some of the "thought process." Claude was trying to figure out how to let me down gently. 

Claude spit out a short book for a prompt... Attached. 

-j

Think carefully

Sent from Proton Mail for Android.

-------- Original Message --------

To unsubscribe from this group and stop receiving emails from it, send an email to 3d-printing-tips--...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/3d-printing-tips--tricks/56b2c301.48e1.19ebeebe8db.Webtop.102%40optimum.net.

[Kurt The 3D Printer GUY!!]

I am Def going to have to Read your Attachment later today. 

And - by the way - you KNOW that Andy called you out SEVERAL Times in the latest podcast. You're now almost as famous as my Doppelgänger claims to be - when he gets the Shout Outs!!!

-K

[Jody Harris]

On the very first day of this, Feature Creep starts sinking in it's deadly talons.

-j

Think carefully

Sent from Proton Mail for Android.

-------- Original Message --------
On Friday, 06/12/26 at 21:00 vr...@optonline.net wrote:

To unsubscribe from this group and stop receiving emails from it, send an email to 3d-printing-tips--...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/3d-printing-tips--tricks/56b2c301.48e1.19ebeebe8db.Webtop.102%40optimum.net.

[Kurt The 3D Printer GUY!!]

Jody - I am SO Frustrated!!!

I went into work today - with the Sole Purposes of using my Claude Pro subscription to try and use Fable 5. Especially since, just recently - when using Claude - it was pushing me to use Fable 5. 

So - I start a new chat session in Claude Code desktop app - switch it to Fable 5 - create a LONG Prompt - only to have it immediately rejected and claim I CAN NOT Use Fable 5!

I even went back to a web browser tab with Claude AI webpage access open - where I was doing a lot of Other work - and I can NOT even access Fable 5 on there!

WHAT GIVES???

How in the HECK were YOU able to do it???

HELP ME!!!

And - FYI - my prompt and request is Way, WAY Simpler than what you are trying to do. But, yes - it's to generate a 3D object - but, generate code that would allow me to create and run a program and tweak options - to produce Many variations of the base 3D Model design to be 3D Printed!

Ugh...

-K

[Jody Harris]

When I logged into the app on my phone this morning, I saw that Fab5 was unavailable. 

I saw a headline this morning that the US government has restricted Fab5 and Mythos5 from foreign users. My suspicion is that Anthropic is working on access control and just removed access to almost everyone while they work that out.

Think carefully

Sent from Proton Mail for Android.

Think carefully

Sent from Proton Mail for Android.

-------- Original Message --------

--
You received this message because you are subscribed to the Google Groups "3D Printing Tips and Tricks" group.
To unsubscribe from this group and stop receiving emails from it, send an email to 3d-printing-tips--...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/3d-printing-tips--tricks/11d5be80-b998-4610-9459-91f6f9623da6n%40googlegroups.com.

[Kurt]

This morning I restarted Claude Code - upon doing so I saw the message they presented to me - which I highlighted in Red:

Upon clicking on the "Learn more" - the following is what appeared in a webpage - essentially confirming what you stated Jody:

Announcements

Statement on the US government directive to suspend access to Fable 5 and Mythos 5

Jun 12, 2026

The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees. The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance. Access to all other Anthropic models will not be affected.

We received the directive from the government today at 5:21pm (ET). The letter did not provide specific details of its national security concern. Our understanding is that the government believes it has become aware of a method of bypassing, or “jailbreaking” Fable 5. We reviewed a demonstration of this specific technique being used to identify a small number of previously known, minor vulnerabilities. These vulnerabilities all appear relatively simple, and we have found that other publicly-available models are able to discover them as well without requiring a bypass.

Anthropic’s posture with respect to Fable’s safeguards, as laid out in our launch blog post, is the following:

• We have instituted strong safeguards that greatly reduce the likelihood that Fable is misused for tasks related to cybersecurity (among others). In fact, our safeguards are so strong that many users have complained that they are overly broad.
• In the weeks leading up to the launch of Fable, Anthropic worked with the US government, the UK AISI, multiple private third-party organizations and internal teams to red-team Fable’s safeguards for thousands of hours in total.
• These tests showed that Fable’s safeguards are substantially more effective than those of any previously deployed model.
• No testers have yet been able to find a universal jailbreak—a jailbreak method that can very broadly bypass the model’s safeguards, unblocking a wide range of cyber capabilities.
• We suspect that perfect jailbreak resistance is not currently possible for any model provider. Every safeguard used in the industry is vulnerable to non-universal jailbreaks (which can elicit some cyber information in specific circumstances), and it is likely that universal jailbreaks will eventually be found in the future. We stated this clearly when we released Fable 5.
• Given that perfect jailbreak resistance does not appear to be possible today, Anthropic adopted a defense in depth strategy with Fable 5. We aimed to make jailbreaks either narrow (in the case of non-universal jailbreaks) or very expensive to produce (in the case of universal jailbreaks), and to combine this with thorough monitoring to quickly detect and shut down any successful attacks. This is also why Anthropic has required 30-day retention of customer data with Fable—a policy change that carries real costs for us with customers, but that allows us to research and mitigate jailbreaks.
• We stand by this defense in depth strategy. It reduces the risks posed by Fable, making them comparable to the risks of existing models already deployed across the industry.
• We have not even received a disclosure of a concerning non-universal potential jailbreak that led to a harmful result. The potential jailbreaks that have been disclosed to us are either entirely benign responses or are minor findings that provide no Mythos-specific uplift.

To date, the government has only given us verbal evidence of a potential narrow, non-universal jailbreak, which essentially consists of asking the model to read a specific codebase and fix any software flaws. Our understanding is that one potential jailbreak was shared with the government. We have reviewed a report that we believe is the basis of the government's directive and validated that the level of capability displayed there is widely available from other models (including OpenAI’s GPT-5.5), and is used every day by the defenders who keep systems safe. We will share more details over the next 24 hours.

We are complying with the government’s legal directive and are removing access to Fable 5 and Mythos 5 for all users. However, we disagree that the finding of a narrow potential jailbreak should be cause for recalling a commercial model deployed to hundreds of millions of people. If this standard was applied across the industry, we believe it would essentially halt all new model deployments for all frontier model providers.

As we have stated publicly, we believe the government should have the ability to block unsafe deployments, as part of a statutory process that is transparent, fair, clear, and grounded in technical facts. This action does not adhere to those principles.

We apologize for this disruption to our customers. We believe this is a misunderstanding and are working to restore access as soon as possible.

To view this discussion visit https://groups.google.com/d/msgid/3d-printing-tips--tricks/14yPNpHOxqWPCuDWa6urpr0Sroa7O3FNhk569VitAr8TrCPmpICuxWFP2hUgwr6LQpZwq1O51Nw1u4xFEfL6Rw5q7LYRpdoB3hVCUZ-RZjw%3D%40harrisdev.com.

[Kurt The 3D Printer GUY!!]

Jody - I got GREAT News yesterday - as Fable 5 is Now available again!!!

-K

On Friday, June 12, 2026 at 8:04:14 PM UTC-4 Jody Harris wrote: