Knowtate

Empowering Research Excellence
Knowtate is a scientific research assistant app offering a suite of features for literature reading, management, and intelligent QA with AI. Embrace the power of centralized document control, seamless reading, and intelligent QA with AI of your document—all within a single, streamlined application.

Note

You can try Knowtate online at: https://knowtate.online or deploy it on your own local enviroment by following the instructions of Quick Start Part.

Features

Literature Management and Reading

Upload and manage your research papers with a central repository backed by object storage service. This service can be self-hostable or various cloud storage solutions like AWS S3. Enjoy an online PDF reader that enhances your literature reading experience.

Diverse Note-Taking Styles:

Knowtate caters to your note-taking preferences with three unique styles, each designed to complement different aspects of your research workflow.

Notion-Styled Rich Text Editor:

For those who appreciate the flexibility and aesthetic of Notion, our rich text editor offers similar functionalities. Utilize slash commands (/) to seamlessly switch between text block styles, supporting multi-level headings, tables, ordered and unordered lists, images, and more.

Markdown Editor:

For the markdown aficionados, the Markdown Editor provides a straightforward, text-focused interface. This allows for efficient writing and formatting with the simplicity and power of markdown syntax—perfect for those who prefer keyboard-centric controls and clean, exportable content.

Excalidraw Whiteboard:

Unleash your creativity with the Excalidraw whiteboard, a digital canvas for freeform drawing, diagramming, and visual brainstorming. This tool is excellent for creating and sharing quick sketches, complex diagrams, or visual notes that complement your research.

Intelligent Paper Q&A:

Powered by a RAG-based intelligent Q&A system, Knowtate allows you to inquire about any aspect of your papers. Customize your dialogue models, vectorization models, and vector databases to suit your needs. Currently supporting OpenAI ChatGPT and OpenAI Embeddings for insightful interactions.

More Features on the Horizon:

Development is ongoing for additional features to enhance your research experience even further.

Quick Start

Run Locally

1. Clone the Repository Locally

# Navigate to your desired path
cd /your/path/
git clone https://github.com/tsmotlp/knowtate.git

2. Configure Environment Variables

cd /your/path/knowtate
cp .env.example .env
# Change `.env.example` to `.env` and update the following variables in `.env`:

# Deployment used by `npx convex dev`
CONVEX_DEPLOYMENT=
NEXT_PUBLIC_CONVEX_URL=
NEXT_PUBLIC_CLERK_PUBLISHABLE_KEY=
CLERK_SECRET_KEY=

MINIO_ENDPOINT=
MINIO_ACCESS_KEY=
MINIO_SECRET_KEY=

OPENAI_API_KEY=
PROXY_URL=

Knowtate uses convex as backend database to store file meta data, please visit convex website to get CONVEX_DEPLOYMENT and NEXT_PUBLIC_CONVEX_URL

Knowtate uses clerk for authentication, please visit clerk website ot get NEXT_PUBLIC_CLERK_PUBLISHABLE_KEY and CLERK_SECRET_KEY

For Windows users, find your network proxy address in Settings -> Network and Internet -> Manual proxy setup -> Edit. Enter the server information in the format http://localhost:10077 as your PROXY_URL.

For Linux users, determine your network proxy address with:

echo $HTTP_PROXY

3. Install and run minio

Download MinIO server binaries compatible with your OS from the MinIO Download Page.

For Windows users, you can open Powershell and enter the following commands:

PS> Invoke-WebRequest -Uri "https://dl.min.io/server/minio/release/windows-amd64/minio.exe" -OutFile "C:\minio.exe"
PS> setx MINIO_ROOT_USER admin
PS> setx MINIO_ROOT_PASSWORD password
PS> C:\minio.exe server F:\Data --console-address ":9001"

For Linux users, you can open Terminal and enter the following commands:

wget https://dl.min.io/server/minio/release/linux-amd64/minio
chmod +x minio
MINIO_ROOT_USER=admin MINIO_ROOT_PASSWORD=password ./minio server /mnt/data --console-address ":9001"

Similar to the MacOS users:

curl --progress-bar -O https://dl.min.io/server/minio/release/darwin-amd64/minio
chmod +x minio
MINIO_ROOT_USER=admin MINIO_ROOT_PASSWORD=password ./minio server /mnt/data --console-address ":9001"

After setting up, MinIO server listens on port :9000, and you can manage your storage through http://localhost:9001.

4. Install Depedencies:

> cd  path/to/knowtate
> npm install # or pnpm install or yarn

5. Run Knowtate:

> cd  path/to/knowtate
> npm run dev # or pnpm run dev or yarn dev

Access Knowtate at http://localhost:3000 and delve into your personalized academic research experience.

Run with Docker

1. Install Docker

For Windows: Download Docker Desktop and follow the official guide or search for an installation tutorial.

2. Clone the Repository Locally

# Navigate to your desired path
cd /your/path/
git clone https://github.com/tsmotlp/knowtate.git

3. Configure Environment Variables

cd /your/path/knowtate
cp .env.example .env
# Change `.env.example` to `.env` and update the following variables in `.env`:

# Deployment used by `npx convex dev`
CONVEX_DEPLOYMENT=
NEXT_PUBLIC_CONVEX_URL=
NEXT_PUBLIC_CLERK_PUBLISHABLE_KEY=
CLERK_SECRET_KEY=

MINIO_ENDPOINT=
MINIO_ACCESS_KEY=
MINIO_SECRET_KEY=

OPENAI_API_KEY=
PROXY_URL=

4. Build the Image

docker-compose build --no-cache

5. Start the Application

docker-compose up

6. Stop the Application

docker-compose down

Copyright and License

The contents of this repository are licensed under the MIT License. We respect and adhere to the copyright and usage permissions of each paper and resource.

Contact Information

For any questions or suggestions, please contact us through the following:

• Submit an Issue
• Email: yqhuang2912@gmail.com