# Definable Knowledge Bases — Your Private Second Brain

> RAG-powered private knowledge layer. Ingest PDFs, databases, URLs, and Git repos. Every answer cites its source. Never used to train models.

## What it does

Definable Knowledge Bases is a private RAG (Retrieval-Augmented Generation) layer that lets you ground AI replies in your own documents and data. Build a queryable "second brain" from your company's information — PDFs, SQL databases, URLs, Git repositories, Google Drive, Notion, Confluence — and ask questions in natural language with citations on every answer.

## Ingestion sources

- **Documents**: PDF, DOCX, TXT, Markdown, HTML
- **SQL databases**: Postgres, MySQL, BigQuery, Snowflake
- **URLs**: with scheduled re-crawls to keep content fresh
- **Git repositories**: public or private, all major languages
- **SaaS sources**: Google Drive, Notion, Confluence, Zendesk
- **Communication**: Slack channels, email archives

## How it works

1. Connect a data source via OAuth or upload files directly.
2. Definable chunks, embeds, and indexes your content with the embedding model of your choice.
3. Ask questions in natural language — the Assistant retrieves relevant chunks and grounds its reply.
4. Every answer includes inline citations linking back to the source.

## Privacy & security

- **Tenant-isolated** — your knowledge base is never queryable by other Definable customers.
- **Never trained on** — your documents are not used to train any model, ours or third-party.
- **Encrypted at rest** with per-customer encryption keys.
- **Per-document ACLs** honour your existing source-system permissions.
- **Bring-your-own embedding model** — OpenAI, Cohere, Voyage, Jina, or open-source.
- **Regional hosting**: India (Mumbai), US (Virginia), EU (Frankfurt).

## Use cases

- Internal company knowledge base — onboarding, policies, procedures.
- Customer support — ground replies in your help center and product docs.
- Sales enablement — quickly pull product specs, pricing, competitive intel.
- Engineering docs — query your codebase, design docs, and runbooks in natural language.
- Research and analysis — synthesize across hundreds of PDFs or reports.

## Pricing

- Starter (free): 1 knowledge base
- Creator ($19/month): 5 knowledge bases
- Pro ($49/seat/month): unlimited knowledge bases + team sharing
- Enterprise (custom): VPC, on-prem, BYO-key for embedding model

## Related

- HTML page: https://definable.ai/knowledge/
- Pricing: https://definable.ai/pricing/
- Glossary — What is RAG?: https://definable.ai/glossary/rag-retrieval-augmented-generation/
- Glossary — What is a knowledge base?: https://definable.ai/glossary/knowledge-base/
- Use case — Build a company knowledge base: https://definable.ai/use-cases/build-company-knowledge-base/