Python SDK

The official Python library for rNet OAuth authentication and authenticated AI calls.

PyPI: rnet-oauth

Installation

bash
pip install rnet-oauth

1. Initialize the Clients

python
from rnet_oauth import RNetAuth, RNetAi

auth = RNetAuth(
    client_id='your-client-id',
    client_secret='your-client-secret',
    redirect_uri='your-redirect-uri'
)
ai = RNetAi()

2. Generate Authorization URL (OAuth2 PKCE)

python
pkce = auth.generate_pkce()
auth_url = auth.get_authorization_url(pkce['challenge'], state='optional-state')

3. Exchange Code for Tokens

python
tokens = auth.exchange_code_for_token(code, pkce['verifier'])
access_token = tokens['access_token']
refresh_token = tokens['refresh_token']

4. Refresh Tokens

python
refreshed_tokens = auth.refresh_access_token(refresh_token)
new_access_token = refreshed_tokens['access_token']

5. Get User Info

python
user_info = auth.get_user_info(access_token)
print(user_info['email'])
print(user_info['name'])

The UserInfo response comes from RNet's /userinfo endpoint and may include: sub, email, email_verified, name, preferred_username, user_id, role, and status.

6. Chat with AI

python
response = ai.chat({
    "contents": [
        {"role": "user", "parts": [{"text": "Hello!"}]}
    ]
}, access_token, "gemini-2.5-flash-lite")

7. Streaming AI Response (Untested)

python
for chunk in ai.chat_stream(payload, access_token, "gemini-2.5-flash-lite"):
    print(chunk)

8. File Upload (Untested)

python
with open("document.pdf", "rb") as f:
    file_buffer = f.read()

# Upload to Gemini
gemini_upload = ai.gemini_file_upload(access_token, "gemini-2.5-flash-lite", file_buffer, "application/pdf", "document.pdf")
print(gemini_upload['fileReference']) # Use this in chat payload

# Upload to OpenAI
openai_upload = ai.openai_file_upload(access_token, "gpt-4o", file_buffer, "application/pdf", "document.pdf")

9. File Deletion (Untested)

python
# Gemini files auto-delete after 48 hours, so there is no delete method.
# Delete an OpenAI file:
ai.openai_file_delete(access_token, "gpt-4o", openai_upload['fileReference'])

10. AI Chat with File & Tools (Untested)

python
payload = {
    "contents": [
        {
            "role": "user",
            "parts": [
                {"text": "Based on this document, what is my name? Also search the web for the current weather in London."},
                {"fileData": {"fileUri": gemini_upload['fileReference'], "mimeType": gemini_upload['mimeType']}}
            ]
        }
    ],
    "tools": [
        {"googleSearch": {}} # Enable Google Search tool
    ]
}

response = ai.chat(payload, access_token, "gemini-2.5-flash-lite")
print(response)

Repository

View the Python SDK on GitHub.

Quick Reference Links