GPT Project Insight

• Project setup
• Run Electron app
• Generate using CLI
• Config and Settings
• Examples

An engine for generating recursive documentation and using it to work on a project with ChatGPT. To use this application, you need to have an API Key for the Open AI API

How it works

Documentation

Insight

Project setup

npm install

Run Electron app

npm run electron:serve

Open By deeplink

gptpi://index?dir=C:/path/to/project&config=C:/path/to/config.json&result=C:/path/to/result.json&prompt=PROMPT

Generate using CLI

npx

npx tsx .\engine\cli.ts YOUR_API_KEY C:/directory/

help

npx tsx .\engine\cli.ts --help

Build the engine and run

tsc ./engine/cli.ts --outDir ./out_engine

node .\out_engine\engine\cli.js YOUR_API_KEY C:/directory/

help

node .\out_engine\engine\cli.js --help

Config and options

Config:

You can load config from CLI using --config option

Config loads automatically from file "SELECTED_DIR/docs.ai.config.json" in the electron app

Options:

• --maxTokens the maximum number of tokens that the model can accept
• --bytesPerToken approximate number of bytes in one token
• --maxQueries Maximum number of requests simultaneously
• --outFile The file to write the result
• --config The file to read the config from
• --maxTokensFile The max tokens values for files --maxTokensDir The max tokens values for directories
• --bytesPerToken approximate number of bytes in one token
• --temperature The temperature of the model
• --excludes The pattern to exclude files. Example: dir1,dir2,file3,*.png
• --model The model to use
• -h, --help display help for command

Examples

Prompt:

Make the footer sticky

Model: GPT-4

Prompt:

How to install the project?

Model: GPT-4

Prompt:

How to create a script: npm run build_engine

Model: GPT-3.5