alexisrobert / dust

Design and Deploy Large Language Model Apps

Home Page:https://dust.tt

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Twitter

Dust

Generative Models App Specification and Execution Engine.

Command-line Client Quick Start

  • Install Rust
  • Clone this repository ($DUST_PATH represents the path to the dust repository).
  • cd $DUST_PATH/core
  • cargo build --release --bin dust
  • alias dust="$DUST_PATH/core/target/release/dust"
  • dust init new_test_project
  • cd new_test_project
  • dust help

See the video demo for an explanation of the commands.

Run Dust Web App Locally Quick Start

Start core API on SQLite

  • cd $DUST_PATH/core
  • cargo run --bin dust-api

Start front on SQLite

  • cd $DUST_PATH/front
  • Create .env.local with the following content:
URL=http://localhost:3000

NEXTAUTH_URL=$URL
NEXTAUTH_SECRET=25da29db5485fcd27d2f671c5e28d5df8f325e9470c36cb6bf0a9d19c662255a

DATABASE_URI=sqlite:front_store.sqlite

GITHUB_ID=47ec837fcbac1c6f1690
GITHUB_SECRET=d59af0da231fc697d7dd8d3002ed1d9765e11c7f

DUST_API=http://localhost:3001

THUM_IO_KEY=0-Foo
GA_TRACKING_ID=Foo
  • Run DATABASE_URI=sqlite:front_store.sqlite ./init/init.sh
  • Run npm install
  • Start server with npm run dev

Questions / Help

https://discord.gg/8NJR3zQU5X

About

Design and Deploy Large Language Model Apps

https://dust.tt

License:MIT License


Languages

Language:JavaScript 54.2%Language:Rust 45.5%Language:Shell 0.1%Language:PEG.js 0.1%Language:CSS 0.0%