✦ SEE-SOL-LAB AI COMPANION ATLAS
Open sourceOriginalMemorySelf-host / OpenAI API

EXTRACT · CONSOLIDATE · CACHE

Pawwake

PAWWAKE · @GARAN0613

Add a layer of memory, conversation lines, and partitioned caching in front of any OpenAI compatible client.

Pawwake sits between the client and the model server, automatically retrieving and injecting relevant memories, and then extracting candidate content from new conversations. It manages both session summaries and cache partitions, reducing long-term chat costs.

Project address (can be copied to AI):https://github.com/garan0613/pawwake

PROJECT INTRO

Project introduction

Pawwake (formerly AI Memory Gateway) is an OpenAI compatible forwarding gateway. It provides three-layer memory of fragments, events, and core, and supports automatic extraction, AI sorting, manual merging, source withdrawal, and keyword + vector hybrid search.

The project also manages cross-platform conversation lines, history records, token statistics and A/B partition cache, and separates authentication for program API and Dashboard login. The client only needs to change the API address to access.

  • Suitable for:People who use clients such as Kelivo, ChatBox, NextChat, etc. and want to obtain memory and cache optimization at the same time.
  • Platform:Python Self-host + PostgreSQL; compatible with any OpenAI format client and model.
  • Usage:First deploy a pure forwarding gateway, and then connect database, memory, vector search and partition cache as needed.
  • License:AGPL-3.0。
◌

COMMENTS & FEEDBACK

Comments and Feedback

Guest message · Will be published after review

People who have used this project can come back and tell newcomers: which platform it ran on, whether there were any pitfalls during the installation, and what the actual experience was like. The current version is in guest mode, and you do not need to register an account to leave a message.

Reading public comments...

0 / 800

SOURCE & CREDIT

Atlas is responsible for the introduction and navigation, and the use still returns to the original author.

We do not mirror project files, nor do we intercept author traffic. For installation, download, version updates and the latest instructions, please refer to the project page provided by the original author.