Documents & content
Gather the texts, files and web pages that matter to your work, within the supported formats.
AI knowledge base
An AI knowledge base that stays under your control.
Bring your documents, your web content and your trade references together in an AI knowledge base you can question and reuse.
Your body of knowledge
An AI knowledge base starts with your sources.
Bases by scope
Organise corpora for a team, a subject or a public-facing use.
Access & uses
Decide who may draw on a base and within which framework it is used.
RAG, plainlyAn AI knowledge base searches your sources
An AI knowledge base searches your sources
before it words its answer.
RAG combines a search across a corpus with the generation of an answer. It brings specific context to the AI, without retraining the model on the whole of your knowledge.
You ask a question
A concrete request is worded in your assistant or your application.
The useful context is retrieved
The system picks the relevant passages from the base you have authorised.
The model drafts the answer
What was retrieved serves as context. The answer still needs checking.
What matters for your data: with an external model, extracts may be sent to it to produce the answer. Storing the base locally is therefore not enough, on its own, to guarantee entirely local processing. Understand the path your data takes
A living foundationAn AI knowledge base is chosen
An AI knowledge base is chosen
and tended.
Start with references you can trust, and name the people responsible for keeping them current.
- A legible scope
Avoid mixing public content with confidential files.
- Current references
Check the versions and remove information that has gone out of date.
- Tested questions
Use the teams’ real requests to spot what is missing.
DocumentsWebsitesTrade knowledgeConnected tools
Your secured knowledge bases
Chosen sources · Defined access · Reusable knowledge
Inside Narratheque
AI for teams, assistants, agents, chatbot
In your tools
Connected applications and automations
The model adapts to the use. Your knowledge stays organised.
Frequently asked questions
What you want to know about an AI knowledge base.
Is a knowledge base just a folder of files?
It gathers sources organised so that AI uses can draw on them. The quality of the corpus, how it is prepared and the access rules bear directly on how relevant the results are.
Can we create a public base and an internal one?
Yes, and those scopes must be kept apart. A base meant for a public chatbot should hold only information that may be given to visitors.
Does RAG remove the AI's mistakes?
No. It brings references but does not guarantee a correct answer. Sources and answers must be checked, particularly for decisions that commit you.
Let us start simplyYour knowledge already has value.
Your knowledge already has value.
Give it new uses.
Let us start from your data, your teams and a first concrete need.
