Yet another ChatGPT UI, but more fit and svelte, and with a secure backend.

Backend/Frontend Sveltekit app to have multiple conversations with chatbots. It connects to the OpenAI API.

• Sveltekit
• Skeleton
• Prisma
• OpenAI API

• Deploy in your own server
• Safely store your API Key in the server
• Run locally in your machine
• Configurable amount of messages to send
• Configurable system message to guide the assistant
• Prompt templates: Speed up your questions with predefined templates
• Prompt template parameters: Add parameters to the template and change them easily for
• Start agents with predefined system message and prompt templates
• Save all the messages in your database, continue your conversations later
• Save the system message and prompt templates of conversations in the database
• You can easily change the database provider to whatever you want
• Auth system your needs.
• Streaming messages
• Edit messages
• Create images with Dall-e
• Voice input
• Voice answers
• Share conversations
• Upload your own documents
• Plugins

• Create a .env file with your OpenAI Api key, and the urls of your postgres databases. See .env.supabase.example for reference. Your file should look like this:

SECRET_OPENAI_API_KEY="sk-XXXX"
DIRECT_URL="postgres:https://postgres:[YOUR_PASSWORD]@project.supabase.co:5432/postgres?connect_timeout=300"
DATABASE_URL="postgres:https://postgres:[YOUR_PASSWORD]@project.supabase.co:6543/postgres?pgbouncer=true"

If you want to use sqlite for a local database, check .env.sqlite.example, your file should look like this:

SECRET_OPENAI_API_KEY="sk-XXXXXXXX"
DATABASE_URL="file:./database_name.db"

• npm install
• npx prisma migrate dev to start the database and apply the schema
• npm run dev to execute

Add the environment variables to your system.

For serverless(e.g. Vercel) you can use Supabase as database, check this documentation.

Hi! you can find me here @julioandresdev