qbud 0.2.0

A package for managing qBud assistants and chats

qBud Python Client

qbud helps to make interaction with qBud assistants easier. Its aim is to provide a low-code solution for interactions with assistants that you have configured via the UI.

How to Install

Installing is easy:

pip install qbud

We recommend python>=3.9, but the client may run with older versions as well. The only external dependency is the widely used requests library, which is installed when running the above command.

Configuration

Authentication is done with a per-assistant access key, generated from your account.

Variable	Required	Default	Purpose
QBUD_ASSISTANT_ACCESS_KEY	yes*	—	Access key for the assistant. *Optional if you pass access_key=... to Assistant() directly.
QBUD_BASE_URL	no	https://api.qbud.ai	Override the API host (e.g. for staging).

Quickstart

1. Connect to an assistant

Make sure you have built an assistant via our UI. When you're editing the assistant there, copy the ID from the address bar (after /assistants/) and pair it with your assistant access key:

from qbud import Assistant

assistant = Assistant("<your-assistant-id>", access_key="<your-access-key>")
# or, with QBUD_ASSISTANT_ACCESS_KEY set in the environment:
assistant = Assistant("<your-assistant-id>")

2. Send messages

Create a chat with assistant.create_chat() and call send_message(...) on it:

chat = assistant.create_chat()
response = chat.send_message("Hi!")
print(response)
> {"content": "Hi there!", "role": "assistant"}

Saving and loading chats

By design, chat message history cannot be retrieved from our API. If you need history across sessions, save chats locally:

from qbud import Chat

# save a Chat object as JSON (message history only — the access key is
# never written to disk, since it is a bearer credential for the chat)
chat.save("chat.json")

# ...session terminates

# load a Chat instance from the saved JSON. The result is read-only:
# get_messages() works, but send_message() requires a live chat created
# via Assistant.create_chat().
chat = Chat.load("chat.json")

You can also inspect the in-memory history with get_messages(...):

print(chat.get_messages())
> [{"content": "Hi!", "role": "user"}, {"content": "Hi there!", "role": "assistant"}]