create-vault

Create Ampi-ready SQLite vaults with FTS5 indexes and semantic tables.

19|4|Updated Jan 4, 2026

Search Tags:#semantic search #search #data management #sqlite #vault #document ingestion

Ready to run

npx skills add https://github.com/wunki/amplify --skill create-vault

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: create-vault
Source: https://github.com/wunki/amplify/tree/main/skills/create-vault
Command: npx skills add https://github.com/wunki/amplify --skill create-vault

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

This Skill automates the creation of searchable SQLite databases (vaults) from collections of documents, making information easily retrievable for AI agents.

Core Features & Use Cases

Document Ingestion: Processes various file types (md, txt, docx, doc) and existing SQLite tables.
Vault Creation: Generates SQLite vaults with keyword and semantic search capabilities.
Use Case: You have a folder full of research papers and notes. Use this Skill to create a vault that allows an AI to quickly find specific information across all documents using natural language queries.

Quick Start

Use the create-vault skill to build a vault from the documents in the '/path/to/my/docs' folder, naming the entities 'articles' and 'authors'.

Frequently Asked Questions about create-vault

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I create a searchable SQLite database from a folder of documents?▼

To create a searchable SQLite database from documents, you ingest files like md, txt, and docx into a vault. The process builds FTS5 indexes and optional sparse semantic tables, enabling keyword and semantic search for AI agents.

What is the best way to enable semantic search across my research notes for an AI agent?▼

Enabling semantic search across research notes involves building an SQLite vault with optional sparse semantic tables. This generates manifest metadata that allows AI agents to quickly retrieve specific information using natural language queries.

Can I build a full text search index from existing SQLite tables?▼

Yes, you can build a full text search index from existing SQLite tables. The vault creation process ingests existing tables and generates FTS5 indexes alongside manifest metadata to facilitate keyword retrieval.

Does the vault creation process support ingesting Microsoft Word documents?▼

The vault creation process supports ingesting Microsoft Word documents. It processes various file types including docx and doc files, adding them to the SQLite database to be indexed for keyword and semantic search.

What file formats are supported for document ingestion into an SQLite vault?▼

Supported file formats for document ingestion include md, txt, docx, and doc files. These documents are processed into SQLite vaults to create FTS5 indexes and optional sparse semantic tables for AI agent retrieval.