Skip to content

Documentation

The documentation ships with the library.

In the meantime, here is an honest preview of the integration. Package and method names are not final yet.

Installation

Install the library in the service that already calls your RAG. Your index, your model and your documents stay the same.

TypeScript, with npm
npm install @wyrlo/sdk
Python, with pip
pip install wyrlo

Package names: [TO BE CONFIRMED]

The five checks

The library sits between your retrieval and your answer, and runs five checks on every question.

  1. Vocabulary

    The question is also searched with the terms used in your documents: “refund” is also searched as “withdrawal”.

  2. Reranking

    The passages found by your retrieval are reordered; only the most useful ones are passed to your model.

  3. Injections

    Instructions hidden in the question or in the passages are detected and blocked.

  4. Verification

    Each sentence of the answer is matched to the passage that supports it. A sentence without a source is flagged.

  5. Decision

    A confidence level is computed (certain, probable or uncertain): answer, or say “I don’t know” when the information is missing.

Then the loop runs on its own

Every night, a loop adjusts your organisation’s settings from its own traces. Every week, a report shows what changed.

See how it works

Coming soon

  • Getting started guide
  • Library reference, in TypeScript and Python
  • LangChain and LlamaIndex integrations
  • REST API for other languages