BrandDocs plugin
A plugin with three skills that learns your Word, PowerPoint or Excel template and generates new documents in the same brand style
- 288 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- Runs a Python engine, reads your templates and writes finished documents to disk
- The visual QA setup script installs LibreOffice and Poppler on the system
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add plugins/branddocsDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Runs a Python engine, reads your templates and writes finished documents to disk
- The visual QA setup script installs LibreOffice and Poppler on the system
This is third-party code. Review the repository files before installing.
What it does
BrandDocs consists of three skills: `brand-docx` for Word, `brand-pptx` for PowerPoint and `brand-xlsx` for Excel. You show the agent one branded `.docx`, `.pptx` or `.xlsx` file. The skill extracts the theme colors and fonts, named styles, section order, layouts, logos and tables, and saves them as a brand profile. New documents are built on the original template shell and use only what the template defines. The generator never writes a literal style name, color or font; they come from the profile. The `verify` command rejects a profile that points at an element missing from the template. Excel keeps formulas and the brand's number formats, PowerPoint uses real masters and layouts and keeps embedded charts native. There is no conversion between formats. The engine runs offline without a model, and the model-assisted steps can only name facts already extracted. The first document from a new template takes up to 15 minutes including checks, and later ones come from the saved profile in seconds. With several templates, `extract --blend` folds a second one into a profile, and `compare-profiles` reports drift between two profiles.
Who it is for. For company staff who regularly need reports, decks, workbooks and letters in a strict corporate template.
Good fit when
- You write reports and memos in a corporate Word template without moving text by hand
- Decks must be assembled from the real masters and layouts of your sample, with charts and tables
- Named regions in an Excel workbook get filled while formulas and number formats survive
- The brand book exists only as a finished file and you need a reusable profile from it
- Two template profiles must be compared in CI to catch color and font drift
Not a fit when
- You have no branded template: the skill learns from a finished file and cannot work without a sample
- You need to move a document from one format to another, for example Word to PowerPoint: the project has no conversion
- You want free design outside the template, since generation uses only artifacts of the sample
- You need release-grade stability for PowerPoint and Excel: the author calls the project alpha and rates Word as the most robust
Example request
Use our company-report.docx template and write a quarterly results report from it, saved as q3-report.docxLimitations
The project is alpha, version 0.10.1. Word is the best proven; PowerPoint and Excel share the same engine and are catching up. The profile schema is frozen and profiles keep working across releases.
You need Python 3.10 through 3.13 and the packages from `requirements.txt`: `python-docx`, `python-pptx`, `openpyxl`, `lxml`, `Pillow`. Visual QA uses LibreOffice and Poppler, and Tesseract is optional for OCR. Without them generation runs with the lowest level of checks. `bash scripts/setup_visual_qa.sh` installs everything, and `python scripts/brandkit/cli.py doctor` shows the state.
Install the whole repository: a single skill folder without the shared engine in `scripts/brandkit/` will not work.
How to disable. In Claude Code disable the plugin through `/plugin`. For Codex remove the three symlinks `brand-docx`, `brand-pptx` and `brand-xlsx` from `~/.codex/skills` and, if you like, the `~/.codex/brand-docs` folder. Saved profiles live in the project or the user folder and can be deleted separately.
FAQ
How do I install BrandDocs in Claude Code?
Run `/plugin marketplace add ferdinandobons/brand-docs`, then `/plugin install brand-docs@brand-docs`. The plugin loads all three skills together with the shared engine. Afterwards prepare the Python packages and check the environment with `python scripts/brandkit/cli.py doctor`.
How do I connect BrandDocs to Codex?
Clone the repository into `~/.codex/brand-docs`, create a virtual environment and install the packages from `requirements.txt`. Then symlink `brand-docx`, `brand-pptx` and `brand-xlsx` into `~/.codex/skills`. If the skills do not show up, restart the agent.
What is a brand profile?
It is a file where the `extract` command records the template's facts: theme colors and fonts, styles, structure, layouts, cover anchors, logos and tables. The generator takes its formatting only from there. You can check a profile with `verify` and improve it with `learn`, `propose-overrides` and `refine`.
Can I try it without my own template?
Yes. The repository ships a synthetic sample `examples/templates/branddocs_template.docx` with matching `.pptx` and `.xlsx` files. You can walk through extraction, verification and generation with them.
Do my documents go to the cloud?
No cloud service is needed: everything runs locally. Real templates are excluded from git in the repository. Only your agent's model sees data, when you ask it to write the document text.
Why is the first document slow?
On a new template it runs extraction, optional model comprehension, content writing and visual QA with a possible repair round. Together this is up to 15 minutes. After that the profile is saved and documents take seconds.
Discussion
No questions or reviews yet.
Related
PluginsA skill that typesets resumes, one-pagers, letters, reports and slides from templates into PDF, PNG and PPTX and checks the result before delivery
A CLI that lets an agent create and edit Word, Excel and PowerPoint: a single binary, no Office install, no extra libraries
A CLI and skill through which the agent creates, reads, converts and edits real docx, xlsx and pptx files locally
A skill to convert Markdown to Word and back through Pandoc with ready styling templates and Lua filters
A skill and CLI that turn Markdown into DOCX, PDF, PPTX, XLSX, PNG and about a dozen more formats, all computed locally
A document pipeline skill: anything to markdown, markdown to clean html and publisher-grade docx, and html back to markdown
Comparison with similar tools
Entries that share the main search query, and how they differ.
| Entry | Official | Repository stars | Main difference |
|---|---|---|---|
| BrandDocs (this entry) | no | 288 | |
| Dashi PPT | no | 10,175 | turns a document into an editable HTML deck and does not learn a brand template |
| vkr-docx: thesis formatting per GOST 7.32-2017 | no | 0 | formats thesis chapters to a GOST standard instead of your template |
README in short
The README explains the idea, names the three skills and the shared steps: extract, comprehend, verify, generate. It continues with prerequisites, installation for Claude Code and Codex, a quick start and the project status.