Scribae: An Intelligent CLI Tool for Converting Markdown Notes into SEO Content

Scribae is an open-source command-line tool whose core function is to help content creators automatically convert scattered Markdown research notes into structured SEO article drafts, supporting multilingual translation and metadata generation. Its core philosophy is to maintain the 'human-in-the-loop' principle, allowing creators to review and edit content at each stage while using large language models to automate repetitive work, emphasizing reproducibility and determinism. This article will introduce this tool in detail from aspects such as background, features, and technical architecture.

Content creators face efficiency dilemmas across multiple stages from research notes to SEO articles: steps like ideation, outlining, writing, polishing, and translation are time-consuming, and switching between tools easily breaks the workflow. Many people use Markdown to record notes, but converting them into standard SEO articles is tedious. Scribae was born precisely to solve this pain point, providing a complete workflow from notes to publication.

Derived from the Latin word 'scribae' (ancient scribe/secretary), Scribae is an open-source CLI tool designed specifically for converting local Markdown notes into structured SEO content packages. It adheres to the 'human-in-the-loop' principle, giving creators control over each link; at the same time, it ensures the predictability and consistency of the process through structured prompts, typed outputs, and response validation, distinguishing itself from other content generation tools.

Scribae's end-to-end workflow includes:

1. Idea Generation Phase: Generate a structured list of article candidates from Markdown notes;
2. SEO Brief Generation: Core feature, generating keyword analysis, outlines, FAQs, and metadata based on notes;
3. Article Writing: Generate drafts by combining notes and briefs, with the ability to specify chapters or reference original notes;
4. Feedback Optimization: Generate improvement reports by comparing drafts and briefs to optimize content quality;
5. Metadata Generation: Automatically generate JSON/Frontmatter for CMS publication;
6. Multilingual Translation: Direct translation using MarianMT (supports mutual translation between English and German/Spanish/French/Italian/Portuguese), and non-direct language pairs can use中转 (transit) or the NLLB model.

Scribae is developed in Python with flexible deployment: supports temporary execution via uvx, global installation via uv, installation via pipx, and direct installation via pip. The translation function relies on PyTorch and Hugging Face Transformers (CPU version can be selected to save space). LLM integration uses OpenAI-compatible APIs, defaulting to connecting to the local Ollama service (ensuring data privacy, cost control, and flexible model switching), and can also be configured with commercial APIs like OpenAI/Anthropic.

Applicable scenarios:

• Technical blog authors: Automate the SEO optimization process;
• Content marketing teams: Batch produce multilingual content while maintaining brand consistency;
• Independent creators: Reduce costs while controlling creation;
• SEO professionals: Quickly generate SEO frameworks. Limitations: Translation relies on PyTorch (approximately 2GB in size); some models cannot fully guarantee deterministic outputs; the CLI tool requires a certain level of technical background, so non-technical users face a learning curve.

Scribae represents the direction of AI-assisted content creation: while retaining the core of human creativity, it automates tedious and repetitive work, becoming a 'smart scribe' for creators. With the progress of LLM technology, such tools will become more intelligent and easy to use. For creators, making good use of such tools is the key to enhancing competitiveness.