Independent project / Early development
LearnVault
Organize your knowledge.
Report with evidence.
Planning AI tools that turn scattered notes and PDFs into structured, source-linked reports you can review and refine.
Explore the project
3D concept artwork, not a product screenshot
From scattered material to clear outputA focused first MVP for knowledge organization and reporting. Accessible 3D creation is a longer-term direction.
What we're exploring
Knowledge you can put to work.
The first proposed product helps independent workers and small teams organize their selected materials and prepare evidence-linked reports. These are plans, not released services.
01 / ORGANIZE
Your materials.
A report you can review.
A planned knowledge organization and reporting tool for selected notes and PDFs. Bring related material together, extract key information, and prepare a draft linked to its sources.
- Group materials by topic and find relevant passages
- Extract key points, decisions, and follow-up work
- Review source-linked drafts and flagged evidence gaps
Illustrative reporting workflow
Draft a project update from my selected notes.
A REVIEWABLE DRAFTGroup progress, decisions, and next steps. Link each supported claim to a source passage and flag missing evidence for review.
Planned workflow, not generated output. Claude API integration is not live yet.
02 / MAKE
3D modeling,
without the steep start.
A planned AI-assisted modeling tool for people without specialist experience. Describe an idea in everyday language, then refine it through understandable steps and approachable controls.
- Plain-language guidance for modeling tasks
- Simple controls for shape and proportion
- Reviewable actions that keep the creator in control
Illustrative modeling workflow
“Make a simple chair with a wider seat.”
- Understand the intent
Identify the chair's parts and the requested change.
- Suggest a clear edit
Adjust the seat width while preserving the overall proportions.
- Review and refine
Let the creator inspect the change before continuing.
Product concept. Model generation and editing are not available on this site.
Development approach
Useful first. Understandable always.
01Connect claims to sources
Prioritize supported reporting and distinguish evidence, uncertainty, and suggestions.
02Keep people in control
Design for understandable steps, reviewable drafts, and deliberate actions rather than opaque automation.
03Measure practical value
Evaluate source accuracy, omissions, user corrections, latency, and cost per report against a simple summarization baseline.
Where things stand
An honest start.
The product is in an early planning phase. This website is a project introduction, not a working reporting service or modeling tool.
CURRENTDefine the reporting MVP
Review reusable Obsidian-based work and scope topic grouping, information extraction, and source-linked report drafts.
NEXTIntegrate and evaluate Claude
Plan retrieval, structured output, and controlled tool calls. Measure evidence quality and report cost.
LATERValidate and expand carefully
Plan interviews and small workflow tests. Explore beginner-friendly 3D modeling as a separate, narrow extension.
About the project
Built independently.
Open to conversation.
LearnVault is led by Mincheol Shin, an independent developer and prospective founder based in Seoul, South Korea. The project is not yet incorporated.
Claude is the intended AI foundation. API integration remains planned, and no commercial partnership or program acceptance is claimed.
Contact the foundermin38@learnvaultai.com
Who is LearnVault for?
The initial reporting concept targets independent workers and small teams handling scattered notes and PDFs. This audience will be validated through interviews and testing. The longer-term modeling direction focuses on non-expert creators.
Can I use the product now?
Not yet. This website introduces the project and its intended direction. It does not provide a live reporting tool or modeling editor.
How will personal materials be handled?
The intended workflow uses materials the user selects. Data handling, storage, and consent requirements will be defined and tested before a public release.