The AI Library Python library provides convenient access to the AI Library REST API from any Python 3.8+ application. The library includes type definitions for all request params and response fields, and offers both synchronous and asynchronous clients.
pip install ailibrary
import os
import ailibrary as ai
client = ai.AILibrary(
api_key=os.environ.get("AI_LIBRARY_KEY"),
domain="https://api.ailibrary.ai/" // only required for self-hosted AI Library instances
)
Initialise your agent
sales_agent = client.agent.create(
title = "Sales Agent"
instructions="You are a sales agent trying to qualify a lead. You are receiving this "
)
Add training files
client.files.upload(
files = ['/local/path/to/file.pdf'], //txt, pdf, pptx, docx, xlsx
knowledge_id = sales_agent.knowledge_id
)
Check status of the agent knowledge
print(client.knowledge_base.get_status())
Chat with agent
completion = agent.chat(
messages = [
{
"role": "assistant",
"content": "Hey, are you looking to buy?"
},
{
"role": "user",
"content": "Yes, I want to know more first"
}
]
)