d’BiYOK Lab: A Parallel PagBiOmicS Space for Practical AI-Assisted Productivity Tools

Artificial intelligence is rapidly becoming part of how we write, research, organize information, apply for jobs, clean data, learn new skills, and build professional workflows. But as these tools become more powerful, another question becomes increasingly important: how can we use AI in a way that is practical, transparent, and more respectful of user control?

This is where d’BiYOK Lab begins.

d’BiYOK Lab is a parallel project space within PagBiOmicS, created to explore practical AI-assisted tools beyond the strictly bioinformatics and omics context. While PagBiOmicS remains connected to biodiscovery, scientific training, omics, and applied research, d’BiYOK Lab expands that mindset toward broader productivity applications.

The idea is simple: build focused tools that help people work better with AI, while making the process easier to understand.

What d’BiYOK Means

The name d’BiYOK combines two ideas.

Bi reflects a Biodiscovery-Integrated perspective: the PagBiOmicS habit of connecting science, data, workflows, and applied problem-solving.

BYOK means Bring Your Own Key. In this context, it refers to applications where users can connect their own AI provider API key, such as Gemini, OpenAI, Claude, DeepSeek, GLM, Ollama, or other compatible systems.

Together, d’BiYOK Lab represents a space for AI-assisted applications that are practical, transparent, educational, and user-oriented.

Why Create a Parallel Space?

PagBiOmicS started from a scientific and bioinformatics-oriented foundation. But many of the same principles that make scientific workflows useful also apply to broader productivity:

- structure messy information,

- make outputs editable,

- keep workflows reproducible,

- help users understand what the tool is doing,

- reduce unnecessary dependence on opaque platforms,

- and treat AI as an assistant, not as a black box.

d’BiYOK Lab allows PagBiOmicS to explore these ideas in a wider setting.

The goal is not only to build tools for researchers. The goal is to support professionals, students, entrepreneurs, educators, job seekers, and everyday users who want practical AI workflows without losing control of their documents, keys, or process.

Local-First and BYOK as Design Principles

Some AI workflows are convenient in the cloud. Others benefit from a local-first approach.

A local-first app can run on the user’s own computer, often through a local browser address such as localhost, while keeping files, exports, and databases on the user’s machine.

The BYOK model complements this idea.

Instead of forcing every user through a single centralized API account, a BYOK tool lets the user connect their own AI provider. This makes costs, limits, and model choice more transparent. It also helps users learn how AI tools actually work.

This does not remove every privacy concern. Users still need to understand each provider’s terms and handle API keys responsibly. But it creates a more educational and user-controlled model than many closed workflows.

Current Application: JobApp AI Assistant

The current application presented under d’BiYOK Lab is JobApp AI Assistant.

JobApp AI Assistant is a CV parser, job matcher, and application exporter. It helps users adapt CV and application materials to specific job opportunities in a more structured way.

Instead of simply asking an AI model to “rewrite my CV”, the app encourages a more careful process:

- import a CV,

- parse it into editable sections,

- select only the experience, skills, projects, and publications that matter for a specific job,

- paste or fetch a job description,

- generate tailored CV bullets, cover letter drafts, and interview preparation notes,

- export the result as editable files.

This matters because job applications are personal and strategic. A CV contains professional history, motivations, skills, and sometimes unpublished or sensitive work. A structured workflow gives users more control while still benefiting from AI assistance.

Two Ways to Use JobApp AI Assistant

JobApp AI Assistant currently has two routes.

1. Local Windows App

The local Windows app is the recommended option for repeated work, longer documents, local exports, and users who prefer not to paste an API key into a browser page.

It runs on the user’s own computer, opens through localhost, and keeps generated files locally.

Useful links:

- GitHub repository: https://github.com/danydguezperez/JobApp-AI-Assistant

- Download Windows app: https://github.com/danydguezperez/JobApp-AI-Assistant/releases/download/v0.1.0/JobApp-AI-Assistant-Windows.exe 

- Download guide: https://github.com/danydguezperez/JobApp-AI-Assistant/releases/download/v0.1.0/JobApp-AI-Assistant-README.pdf 

2. Web App Lite

The Web App Lite is an experimental browser version hosted through GitHub Pages.

It can:

- import CV files directly in the browser,

- parse PDF, TXT, Markdown, or JSON,

- split the CV into editable sections,

- allow users to activate or deactivate sections before generation,

- download a focused CV as Markdown,

- paste or attempt to fetch a job description from a URL,

- generate an application draft with Gemini using the user’s own API key.

Try Web App Lite:

https://danydguezperez.github.io/JobApp-AI-Assistant/jobapp-web-lite.html 

The important caution is the API key. Because the user pastes the key into a browser page, the recommended workflow is:

1. Create a temporary, restricted, or low-limit API key.

2. Use it for the test.

3. Clear the key field from the page.

4. Revoke or delete the key in the provider account after testing.

PagBiOmicS does not store the key in this static Web App Lite. However, browser environments can still carry risks, including extensions, shared computers, injected scripts, autofill, and careless reuse of keys.

Gemini API key guide:

https://ai.google.dev/gemini-api/docs/api-key 

What JobApp AI Assistant Can Do

JobApp AI Assistant currently supports workflows such as:

- CV import and parsing,

- structured editing of CV sections,

- role-focused filtering before AI generation,

- job description matching,

- tailored CV bullet generation,

- cover letter drafting,

- interview question preparation,

- local application history in the desktop app,

- export to editable formats such as DOCX, PDF, Markdown, and JSON,

- connection to AI providers through user-configured API keys or compatible local models.

The desktop app remains the stronger option for local work. The Web App Lite is useful for browser-based tests and lightweight workflows.

A Practical Educational Layer

d’BiYOK Lab is not only about releasing tools. It is also about helping users understand AI infrastructure in simple terms.

Many people use AI daily without knowing what an API key is, where their data goes, what a model provider can see, or why local execution may be better for some workflows.

Each d’BiYOK app should therefore help answer practical questions:

- What is being sent to the AI model?

- Which provider is being used?

- Where are files stored?

- Can I use my own API key?

- Can I run part of the workflow locally?

- How can I export and keep control of my outputs?

- How should I handle browser-entered API keys safely?

This educational layer is part of the project.

Beyond Research: Wider Productivity Applications

Although the project grows naturally from PagBiOmicS, d’BiYOK Lab is not limited to bioinformatics or academic research.

The same approach can support many types of practical tools:

- literature and document assistants,

- data cleaning utilities,

- job application workflows,

- business planning helpers,

- learning and training apps,

- writing and editing tools,

- local knowledge-base assistants,

- small automation tools for professionals and citizens

Closing Perspective

d’BiYOK Lab is a parallel PagBiOmicS space for exploring practical AI-assisted productivity tools.

It is experimental, but intentional.

The aim is to build small, focused applications that solve real problems, make AI workflows easier to understand, and give users more control over their own data, keys, documents, and outputs.

JobApp AI Assistant is one step in this direction: a practical tool for CV parsing, job matching, and editable application generation.

More tools may follow as this space grows.