Skip to content

Overview

The Private Chatbot API is a software developer kit designed for developers to leverage Private LLMs to upload documents, chat with them and evaluate the answers. There are 3 API endpoints:

  • Upload: Upload documents or specify a ticker for data retrieval and Q&A.

  • Chat: Send a message to the chatbot based on documents previously uploaded.

  • Evaluate: Evaluate predictions on one or multiple documents/text.

Setup

This setup guide is designed to help get your local development environment setup and send your first API request. Throughout this guide, you will learn:

  • How to setup your Private Chatbot account
  • How to install the Private Chatbot Python Pip Package
  • How to send your first Private Chatbot API request

Step 1: Account setup

First, create an Private Chatbot (https://Anote.ai/) account or sign in. Next, navigate to the API key page and "Create new API key". Copy the API key, and make sure to save this somewhere safe and do not share it with anyone.

Step 2: Installing Python Pip Package

To use the Private Chatbot Python library, you need to have Python installed. To ensure you have Python installed, navigate to your Terminal or Command line:

  • MacOS: Open Terminal: You can find it in the Applications folder or search for it using Spotlight (Command + Space).

  • Windows: Open Command Prompt: You can find it by searching "cmd" in the start menu.

Next, enter python3 --version and then press enter, to ensure you see a python version >3.7.

To install the Private Chatbot Python library from the terminal / command line, run: 

pip install -U Anote

Step 3: Sending your first API request

After you have Python configured and set up an API key, the final step is to send a request to the Private Chatbot API using the Python library. To do this, create a file named Anote-test.py using the terminal or an IDE. Download these documents and move them to your local directory.

Now, you have two choices: you can either use the private models (LLaMa or Mistral), or you can use the public models (GPT-4 or Claude).

If you want to use private models, inside the file, copy and paste the example below: 

from anoteai import Anote

api_key = 'INSERT_API_KEY_HERE'
Anote = Anote(api_key, isPrivate=True)

file_paths = ['doc1.pdf', 'doc2.pdf']

upload_result = Anote.upload(task_type="documents", model_type="llama", file_paths=file_paths)
print("output from upload: ", upload_result)
chat_id = upload_result['id']
chat_result = Anote.chat(chat_id, "What is this paper classification performance about?")
print("output from chat: ", chat_result)
message_id = chat_result['message_id']
print("output from evaluate:", Anote.evaluate(message_id))

If you want to use public models, inside the file, copy and paste the example below: 

from anoteai import Anote

api_key = 'INSERT_API_KEY_HERE'
Anote = Anote(api_key, isPrivate=False)

file_paths = ['doc1.pdf', 'doc2.pdf']

upload_result = Anote.upload(task_type="documents", model_type="gpt", file_paths=file_paths)
print("output from upload: ", upload_result)
chat_id = upload_result['id']
chat_result = Anote.chat(chat_id, "What is this paper classification performance about?")
print("output from chat: ", chat_result)
message_id = chat_result['message_id']
print("output from evaluate:", Anote.evaluate(message_id))

If you want to use both private and public models, inside the file, copy and paste the example below: 

from anoteai import Anote

api_key = 'INSERT_API_KEY_HERE'
Anote_private = Anote(api_key, isPrivate=True)
Anote_public= Anote(api_key, isPrivate=False)


file_paths = ['doc1.pdf', 'doc2.pdf']

upload_result = Anote_private.upload(task_type="documents", model_type="llama", file_paths=file_paths)
print("output from upload private: ", upload_result)
chat_id = upload_result['id']
chat_result = Anote_private.chat(chat_id, "What is this paper classification performance about?")
print("output from chat private: ", chat_result)
message_id = chat_result['message_id']
print("output from evaluate private:", Anote_private.evaluate(message_id))

upload_result = Anote_public.upload(task_type="documents", model_type="gpt", file_paths=file_paths)
print("output from upload public: ", upload_result)
chat_id = upload_result['id']
chat_result = Anote_public.chat(chat_id, "What is this paper classification performance about?")
print("output from chat public: ", chat_result)
message_id = chat_result['message_id']
print("output from evaluate public:", Anote_public.evaluate(message_id))

To run the code, enter python Anote-test.py into the terminal. The output should look as follows (may differ slightly depending on whether you have selected private/public):

output from upload:  {'id': 49}
output from chat:  {'answer': 'This paper, "Improving Classification Performance With Human Feedback", is about enhancing the accuracy, recall, and precision of AI models with the help of human feedback', 'message_id': 87, 'sources': ['doc1.pdf', 'Improving Classification Performance With Human Feedback:\n\nLabel a few, we label the rest\n\nEden Chung, Liang Zhang, Katherine Jijo, Thomas Clifford, Natan Vidra\n\nAbstract\n\nIn the realm of artificial intelligence, where a vast majority of data is unstructured']}
output from evaluate: {'answer_relevancy': 0.887419956065742, 'faithfulness': 1.0}