🌊 Wave

     

Cut through the clutter. A knowledge hub for humans and AI agents to find information when you need them. No scrolling through scattered folders, wikis, and documents.

Always up-to date with agents and human moderation.

Features

• Beautiful distraction free Markdown editor with user collaboration
• Content categorization with sharing with users
• Agent memory with virtual file-system support
• Bring your own keys for LLM and S3 storage
• Self hosted and Entangle Cloud host support

Screenshots

Environment Variables

To run this project, you will need to add the following environment variables to your .env file in the frontend and backend

Frontend env variables

Replace with Wave backend url

VITE_API_ENDPOINT="http://localhost:3000

Backend env variables

OPENVIKING_TOKEN="YOUR OPEN VIKIN TOKEN"
DATABASE_URL="POSTGRES DATABASE_URL"
OPENVIKING_MCP_URL="OPENVIKIN MCP ENDPOINT"
OPENVIKING_ENDPOINT="OPEN VIKING ENDPOINT"
JWT_SECRET="xxxxxx"
ACCESS_TOKEN_EXPIRE_MINUTES=60
R2_ACCESS_KEY="S3 BUCKET ACCESS KEY"
R2_ACCESS_SECRET="S3 BUCKET ACCESS SECRET"
R2_ENDPOINT="S3 ENDPOINT"
BUCKET_NAME="S3 BUCKET NAME"
GEMINI_API_KEY="GEMINI API KEY"
MODEL="models/gemini-3.8-flash"

Roadmap

• Wave MCP API Endpoint for agent workflows
• Improving user interface

Architecture

flowchart TD
    A[Frontend] --> B(API - Fast API)
    B -->|Metadata| D[Postgres Database]
    B -->|Agent Memory| C[OpenViking]
    B -->|Agent Workflow| E[Your LLM]
    E --> C
    B -->|Media Storage| G[S3 Storage]
    D --> G 

Loading