Chat with PDF

Introduction

Chat with PDF Tool allows users to upload a PDF, ask questions based on its content, and receive answers using OpenAI model. This is a RAG application implemented using Langchain (opens in a new tab), and Pinecone (opens in a new tab) (for storing document embeddings).

Quickstart Guide

Installation

Clone the repository:
terminal
```
git clone https://github.com/1811-Labs-LLC/BuilderKit-Pro.git [YOUR_APP_NAME]
 
cd [YOUR_APP_NAME]
 
git checkout chat-with-pdf
```
Remove the origin remote to ensure that you can work locally without pushing the changes back to the original repository.
terminal
```
git remote remove origin
```
However, note that after removing the remote, you won't be able to switch between the branches, so you'll need to clone the repository again if you want to work on another branch.
Install dependencies:
terminal
```
npm install
```

Environment Variables:

Copy the required variables from .env.example to .env.local as mentioned and update the values.

terminal

cp .env.example .env.local

Or, manually create a .env.local file in the root directory with the following env variables (make sure to update the values with your actual keys):

.env.local

# Host
NEXT_PUBLIC_APP_URL=<your-app-url>
 
# Supabase
NEXT_PUBLIC_SUPABASE_URL=<your-supabase-url>
NEXT_PUBLIC_SUPABASE_ANON_KEY=<your-supabase-anon-key>
SUPABASE_SERVICE_ROLE_KEY=<your-supabase-service-role-key>
SUPABASE_STORAGE_BUCKET_NAME=<your-supabase-storage-bucket-name>
 
# OpenAI
OPENAI_API_KEY=<your-openai-api-key>
 
# Pinecone
PINECONE_API_KEY=<your-pinecone-api-key>
PINECONE_INDEX_NAME=<your-pinecone-index-name>
 
# Google Analytics
NEXT_PUBLIC_GOOGLE_ANALYTICS_KEY=<your-google-analytics-key>

Setup Supabase

If you have not setup the supabase yet, go through the detailed documentation to set up the supabase for BuilderKit.ai. Make sure that you are creating the user table as mentioned in the supabase setup doc as it is required for the tool.

Create Chat with PDF Table in Supabase:

sql

-- Create a table for Chat with File
create table chat_with_file (
   id uuid not null default uuid_generate_v4 (),
   created_at timestamp with time zone not null default now(),
   user_id uuid null,
   file text not null,
   filename text not null,
   chat_history jsonb null,
   history_metadata text null,
   constraint chat_with_file_pkey primary key (id),
   constraint chat_with_file_user_id_fkey foreign key (user_id) references users (id)
);
 
-- Set up Row Level Security (RLS)
alter table chat_with_file
enable row level security;
 
create policy "Users can insert their own row." on chat_with_file
for insert with check (auth.uid() = user_id);
 
create policy "Users can update own row" on chat_with_file
for update using (auth.uid() = user_id);
 
-- Optional: Add policy to allow users to delete their own row
create policy "Users can read own row" on chat_with_file
for select using (auth.uid() = user_id);

For all the tables, we enable the RLS policy by default with necessary permissions as mentioned in the script.

Sync Supabase Types:

To sync the supabase table schema with your project follow the steps here.

Running the Application

Run the development server:
terminal
```
npm run dev
```
This will start the development server on http://localhost:3000 (opens in a new tab).
Build for production:
terminal
```
npm run build
```
This command compiles the application for production usage.
Start the production server:
terminal
```
npm start
```
This will start the application in production mode.

Additional Scripts

Prepare Husky for Git hooks:
terminal
```
npm run prepare
```
Validate the code with Linting, Formatting & Typecheck:
terminal
```
npm run validate
```

Make the tool publicly accessible by removing the login process

Go to src/app/(dashboard)/layout.tsx and comment the following section from line:15 to line:20 as shown in the image below:
Now, go to src/app/api/chat/route.tsx and comment the following section from line:20 to line:23 as shown in the image below:

These sections reads the user data and checks if the user is signed in or not. By commenting this section it does not check the user validation.

Requirements

Node.js: Download and install from here (opens in a new tab).
Supabase: Check out Supabase Setup Documentation to get all the required details.
OpenAI API Key: Sign up for an API key on OpenAI (opens in a new tab).
Pinecone: Create an account in Pinecone (opens in a new tab) to create an API key and Index name.

MultiLLm ChatGPT Voice Transcription