Does this skill require specific SQLite extensions?

Yes, it requires the FTS5 extension to be enabled in SQLite for keyword indexing. Most modern Python and system SQLite builds include this by default.

Which document formats are supported for ingestion?

The skill natively supports .md, .markdown, .txt, and .docx files. Legacy .doc files are supported via best-effort parsing if system tools like antiword are installed.

What is an Ampi vault?

An Ampi vault is a SQLite database structured with a specific schema contract that allows Ampi tools to perform standardized keyword, semantic, and deep searches.

Can I use domain-specific names for my data?

Absolutely. You can define custom entity names (e.g., 'tickets' and 'customers' instead of 'chunks' and 'docs') to match your specific project domain.

How do I ensure my vault was created correctly?

The skill includes a 'check' command that validates the schema, manifest, and indices, returning a readiness status and detailed error reports if issues are found.

Ampi Vault Builder

Name: Ampi Vault Builder
Author: wunki

bywunki

•

Database Management

Transforms document folders or SQLite tables into searchable, Ampi-compliant database vaults with keyword and semantic indexing.

Ampi Vault Builder is a specialized utility designed to create structured SQLite 'vaults' that integrate seamlessly with Ampi's advanced search and retrieval ecosystem. It automates the process of ingesting documents—including Markdown, TXT, and Word files—and organizes them into a standardized schema featuring FTS5 keyword indexing and optional sparse semantic search. By enforcing a strict search contract and generating a manifest, it ensures your data is immediately ready for deep search and record lookup tools, making it an essential component for developers building Retrieval-Augmented Generation (RAG) workflows or internal knowledge bases.

Key Features

01Standardized schema contract generation (chunks and docs views)

02Integrated validation suite to ensure vault readiness and data integrity

0317 GitHub stars

04FTS5 keyword search and optional sparse semantic indexing

05Customizable domain-specific entity naming for better data modeling

06Automated document ingestion for Markdown, TXT, DOCX, and DOC formats

Use Cases

01Building high-performance retrieval layers for AI agents and RAG applications

02Creating a searchable knowledge base from a folder of technical documentation or research findings

03Standardizing an existing SQLite database for use with Ampi-powered search tools

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add wunki/amplify create-vault

For use in Claude.ai and ChatGPT

Download Skill