reachable

Work In Progress!

Reachable is a tool that enables the utilization of large language models (LLMs) in real-world applications by leveraging embeddings to expand the model's contextual understanding. With Reachable, data such as documentations and codebases can become more accessible and easily understood by LLMs.

Motivation

As the use of LLMs becomes increasingly prevalent in various fields, the limitations of their context window have become a challenge in their effective application. To overcome this challenge, Reachable utilizes embeddings to provide LLMs with almost unlimited memory capacity. By doing so, LLMs can understand and analyze data in a more comprehensive manner, allowing for improved performance and efficacy.

Features

Reachable offers a package and CLI tool that facilitates the creation and management of embeddings for various data types, including text and code. By using Reachable, you can:

• Generate embeddings for your data
• Store embeddings in a database (e.g. pinecone)
• Retrieve embeddings and associated data for use with LLMs
• Perform similarity searches on embeddings to find related data