ChatGPT can help you draft small LISP routines for BricsCAD. Use it as a helper, not as the final judge. The best results come from one clear task, a precise BricsCAD setup, and a careful review. This guide also covers AI LISP coding, ChatGPT AutoLISP, BricsCAD LISP, and cad automation lisp in a simple way.
At NORI CAD, we see BricsCAD as a strong base for controlled, repeatable drafting. It supports LISP-based custom setup and can load and run LSP files. BricsCAD also supports AutoLISP, APPLOAD, and auto load through on_start.lsp and on_doc_load.lsp.
Yes. ChatGPT can draft, explain, and revise LISP routines for BricsCAD. It is most useful for first drafts, command names, comments, and small fixes. It is not a replacement for technical review or testing.
That matters because AI LISP coding is only safe when the code is checked. A routine can select objects, change layers, delete data, save files, or write outside the drawing. Those parts need human review.
BricsCAD is a good home for BricsCAD AutoLISP work because it runs where drafters already work. It can load LSP files and use vl, vlr, vla, and vlax functions. That makes it useful for small jobs like setting a layer, checking text height, or counting objects.
The value is not in big promises. It is in removing small, repeated drafting frictions and reducing variation. Do not expect every online example to behave the same in every setup. Version, operating system, support paths, loaded apps, drawing standards, object types, and user settings can all change the result.
Here at NORI CAD, we recommend treating BricsCAD as the stable CAD platform and ChatGPT as a code helper. That pairing works well when the process is disciplined.
A good routine begins with one clear task. Do not ask one command to do five jobs. If the task has three parts, split it into three routines.
A simple workflow looks like this:
Good early tasks include counting selected objects, setting the current layer, drawing a simple marker, changing text height, or checking whether selected objects are on approved layers.
Higher-risk tasks should wait. These include deleting geometry, batch editing many drawings, renaming layers, overwriting data, saving files, exporting files, importing data, or changing final files. Those jobs can still be done with cad automation lisp, but they need more review.
The prompt acts as the specification for ChatGPT. If it is vague, ChatGPT may make assumptions, which can lead to unclear or unwanted behaviour.
Always include these details:
Be clear about what the routine may change and what it may not change. If the routine saves, renames, deletes, overwrites, exports, or writes external files, say so. If it must not do those things, say that too.
Use a simple format when you ask ChatGPT to act as a LISP routine generator.
I am using BricsCAD [version] on [operating system]. Write a BricsCAD LISP routine saved as an LSP file. The command name must be [command name]. The routine must act on [object types]. The drawing units are [units]. The routine must use or check these layers: [layer details]. The expected result is [result]. If the user selects nothing, the routine must [behaviour]. If the selection contains unsupported objects, the routine must [behaviour]. The routine must not save, rename, delete, overwrite or write external files unless explicitly stated. Add comments explaining the main steps. Keep the routine short and easy to review. Include brief loading and testing instructions for BricsCAD.
A filled example:
I am using BricsCAD V26 on Windows. Write a BricsCAD LISP routine saved as an LSP file. The command name must be C:COUNTLINES. The routine must ask the user to select objects and report how many are lines. The drawing units are millimetres. It must not create geometry or change layers. If the user selects nothing, it must print a clear message and make no drawing changes. If the selection contains unsupported objects, it must ignore them. The routine must not save, rename, delete, overwrite or write external files. Add comments. Keep it short. Include brief loading and testing instructions for BricsCAD.
A starter routine for this prompt would simply count LINE objects and print the total. That is enough to check the workflow without adding risk.
Small routines are easier to read and easier to fix. That is why the best first use of write lisp with chatgpt is a simple one-step command.
Good early BricsCAD LISP tasks:
Wait on bigger jobs until the review method is ready. A routine that purges data, renames layers, overwrites attributes, saves drawings, exports files, imports data, or deletes objects needs more care.
BricsCAD can load and run LSP files. APPLOAD provides a straightforward manual way to load an LSP file into BricsCAD.
A simple loading process is:
After a routine has passed review, BricsCAD can also use on_start.lsp and on_doc_load.lsp for auto load. Use that only for accepted routines. Do not use auto load for a fresh prototype.
A routine should be tested on copies before it is used in real work. One good test is not enough.
Use a staged test set:
Also test user behavior. Press Escape. Select too many objects. Select the wrong objects. Run the command twice. Undo the command if it changes the drawing. Close and reopen the copy. These checks often show problems that a clean demo will hide.
If the routine saves a file, writes data, or edits many drawings, add a higher level of review. Check paths, names, access rights, and backups first.
A routine that works once may still be wrong for a team. That is true when many drafters use it, or when it affects final files.
Before wider use, check these points:
Keep a dated folder for accepted routines. Save the prompt, the output, manual edits, and test notes there. If the routine changes later, keep the older copy until the new one is proven.
The best case for BricsCAD LISP is steady use. One small routine saves a little time. Ten good routines can cut repeat work and reduce variation across drawings.
ChatGPT can help with syntax, command structure, comments, and simple logic. It can speed up AI LISP coding. But the user still has to give the spec and the judgment.
A good team rule set is simple:
This keeps ChatGPT lisp output useful without treating it as safe by default.
At NORI CAD, we provide BricsCAD software, training, and support for teams that want practical custom setup. If you want to use ChatGPT output inside BricsCAD, start with a small routine and build a review process around it.
BricsCAD is a strong platform for controlled drafting automation. It works best when the script is reviewed, tested, and deployed with care.
Question: Should I run a ChatGPT-generated LISP routine in a live BricsCAD drawing?
Short answer: No. Treat the output as a prototype. Load and test it in copy drawings first. Then test it in representative project copies. Use it in live work only after review and testing.
Question: What should I include in the prompt?
Short answer: Include the BricsCAD version, OS, object types, units, layers, command name, expected result, empty selection behavior, wrong-object behavior, and file change rules. The more exact the prompt, the easier the review.
Question: Why start with small, non-destructive routines?
Short answer: They are easier to check, easier to fix, and less risky. Tasks like counting selected lines or listing block names help you learn the process without harming project data.
Question: What is a straightforward way to manually load a prototype LSP file in BricsCAD?
Short answer: Use APPLOAD. Save the reviewed code as a .lsp file, open a copy drawing, load the file, and run the command. Use on_start.lsp or on_doc_load.lsp for automatic loading only after the routine has passed review and testing.
