# AI use cases for Loopcut (Samffprice/loopcut) Repo: [Samffprice/loopcut](https://github.com/Samffprice/loopcut) | Use cases: 4 | Page: https://dedupedai.com/use-cases/samffprice-loopcut/ Loopcut is an open-source AI agent built into a fork of Blender 5.2 that writes and runs Python to carry out technical scene work from plain-language requests. This file holds every use case of this repo. Each one starts with a "## " heading and ends with its URL. - Clean up a messy Blender scene from one message: https://dedupedai.com/use-cases/samffprice-loopcut/clean-up-scene-from-one-message/ - Match a scene's framing to a reference photo: https://dedupedai.com/use-cases/samffprice-loopcut/match-scene-framing-to-reference-photo/ - Build an IK leg rig with mirrored .L and .R names: https://dedupedai.com/use-cases/samffprice-loopcut/build-ik-rig-with-mirrored-naming/ - Build Loopcut from source to change its add-on: https://dedupedai.com/use-cases/samffprice-loopcut/build-loopcut-from-source-to-change-add-on/ ## How to clean up a messy Blender scene from one message with Loopcut Repo: [Samffprice/loopcut](https://github.com/Samffprice/loopcut) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README Rename, sort, fix normals and purge orphans in one request, with each step approved first. **Who it is for:** A 3D artist or technical artist with a cluttered .blend file who needs naming, collections and mesh hygiene fixed before handoff or rendering. **The problem:** Manual cleanup of names, collections, normals and unused data is slow and easy to get wrong across many objects. **The outcome:** The scene is renamed, sorted and cleaned, with a Scene changes card listing what was added, removed and changed. ### Steps 1. Download the latest build from the releases page (https://github.com/Samffprice/loopcut/releases/latest) and install it: drag Loopcut to Applications on macOS, or run the .msi on Windows. 2. On first launch, choose Import Blender 5.2 Settings or Start Fresh, then connect a model in the panel. 3. Open the Loopcut panel with ⌘ L on macOS or Ctrl Alt L on Windows. 4. Type a request such as: Rename everything after what it is, put the lights in their own collection, and apply scale on everything. 5. Approve each code step as it appears, or tell the agent to stop asking for this conversation. 6. Read the Scene changes card, and use Undo all if the result is not what you wanted. ### Prompt to paste into your agent ```text Using the Loopcut add-on in the Samffprice/loopcut repo, ask the agent to rename objects in by what they are, sort lights into their own collection, and purge orphan data. Expect a Scene changes card listing every added, removed and changed item. Verify that the card matches the outliner, and that Undo all restores the scene. ``` ### Caveat The README says the builds are not yet signed, so macOS and Windows will each show a warning on first launch that you must click through. Risk, remote script: Loopcut writes and runs Python inside your open scene, so review each code step before approving it, especially on files you cannot easily recreate. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/samffprice-loopcut/clean-up-scene-from-one-message/ ## How to match a scene's framing to a reference photo with Loopcut Repo: [Samffprice/loopcut](https://github.com/Samffprice/loopcut) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README Attach a reference image and let the agent compare viewport captures and adjust the shot. **Who it is for:** A product or lighting artist who has a reference photo of the look they want and needs the Blender viewport and render setup to match it. **The problem:** Matching camera framing and render settings to a photo by hand means many rounds of adjusting and checking. **The outcome:** The scene framing moves toward the reference, and a render preset is set up at the requested size. ### Steps 1. Install Loopcut and connect a model, following the install steps in the README. 2. Open the panel with ⌘ L on macOS or Ctrl Alt L on Windows. 3. Drop your reference photo onto the panel, or use + image, to attach it to the message. 4. Type a request such as: Match the framing of the reference photo I attached and set up a product-shot render preset at 2048². 5. Let the agent capture the viewport and compare it with the reference, then approve the steps it proposes. 6. Check the final viewport yourself, and use Undo all if the framing does not suit the shot. ### Prompt to paste into your agent ```text Use the Samffprice/loopcut Loopcut agent to frame to match the attached reference image, using capture_viewport and compare_with_reference. Expect an adjusted camera or viewport and a render preset at 2048 pixels. Verify by comparing the final capture with the reference visually before accepting the change. ``` ### Caveat The README describes comparison with a reference as a check the agent makes, not as a guarantee of an exact match, so the final framing still needs your judgement. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/samffprice-loopcut/match-scene-framing-to-reference-photo/ ## How to build an IK leg rig with mirrored .L and .R names with Loopcut Repo: [Samffprice/loopcut](https://github.com/Samffprice/loopcut) | Difficulty: Intermediate | Time: 45 minutes | Basis: From the README Ask for IK chains with pole targets and consistent left and right naming in one request. **Who it is for:** A rigger preparing a character or mechanical model for animation who needs IK chains and consistent side naming set up quickly. **The problem:** Setting up IK chains, pole targets and matching left and right names across many bones is repetitive and error-prone. **The outcome:** The legs have IK chains with pole targets, and the naming follows the .L and .R convention, reported in the Scene changes card. ### Steps 1. Install Loopcut from the releases page and complete first launch, as described in the README. 2. Connect a model, either your own key, a local model, or hosted sign-in. 3. Open the panel with ⌘ L on macOS or Ctrl Alt L on Windows. 4. Type a request such as: Add an IK chain to the legs with pole targets, and mirror the naming to .L/.R. 5. Approve each code step, then review the Scene changes card for the constraints and renamed items. 6. Test the rig by posing it in Blender before you rely on it. ### Prompt to paste into your agent ```text Using the Samffprice/loopcut Loopcut add-on, ask the agent to add IK chains with pole targets to the legs of and mirror bone names to .L and .R. Expect new IK constraints and consistent naming reported in the Scene changes card. Verify with get_object_info that the constraints exist and that each left side name has a matching right side name. ``` ### Caveat The agent writes its own Python for the rig, so a rig that looks correct in the diff still needs a posing test in Blender before animators use it. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/samffprice-loopcut/build-ik-rig-with-mirrored-naming/ ## How to build Loopcut from source to change its add-on with Loopcut Repo: [Samffprice/loopcut](https://github.com/Samffprice/loopcut) | Difficulty: Advanced | Time: 1 hour | Basis: From the README Clone the fork, build the dev app and run the add-on tests to modify Loopcut. **Who it is for:** A developer who wants to change the Loopcut add-on or the fork itself, and needs a working dev build with tests. **The problem:** Changing the add-on or the fork needs a local build that reloads the panel and a way to run tests without opening Blender. **The outcome:** A dev build of Loopcut is installed and the Blender-free tests run. ### Steps 1. Clone the repository from https://github.com/Samffprice/loopcut and open the directory on the main branch. 2. Run `make update` to fetch Blender's libraries from projects.blender.org. 3. Run `ninja -C ../build/lite install` to build the dev app. 4. Edit the add-on in scripts/addons_core/loopcut/, or run loopcut/scripts/dev.sh to reload the panel when you save a file. 5. Run `python3 -m unittest discover -s loopcut/tests` to run the Blender-free tests. ### Prompt to paste into your agent ```text In the Samffprice/loopcut repo, run `make update`, then `ninja -C ../build/lite install`, then `python3 -m unittest discover -s loopcut/tests`. Expect the build to finish and the tests to report a result. Verify that the test run completes without Blender open and report any failing test names. ``` ### Caveat The README gives the build commands but not the system requirements for compiling a Blender fork, so expect a long first build on a machine that already has the needed toolchain. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/samffprice-loopcut/build-loopcut-from-source-to-change-add-on/