Authentication and Data Security: Sidekick's OpenAI Integration and Usage
Sidekick is an application that integrates with OpenAI's API to provide AI-powered features. To use this integration, users must authenticate with their OpenAI account and grant Sidekick access to their API key. The app then uses this key to communicate with the OpenAI API while keeping user data secure. This article examines Sidekick's implementation of OpenAI integration, including how the app handles authentication, privacy, and costs.
To authenticate with OpenAI's API, users must first enable billing on their OpenAI account and create an API key specifically for Sidekick. This process involves navigating to the account settings, selecting "Billing" to set up a paid account, and then generating a new secret key through the API keys section of the dashboard.
Once the API key is created, users copy the key and open the Sidekick app's settings. Here, they select "Add API Key" and paste their newly copied key. This two-step authentication process ensures that Sidekick can communicate with the OpenAI API while maintaining user privacy and security.
Sensitive information, including the OpenAI API key, is stored locally in the user's keychain and never transmitted to Sidekick or any external servers. This local storage approach prevents the app developer from accessing or retaining any authentication credentials.
For analytics and crash reporting, Sidekick utilizes Firebase, which collects anonymous data to improve app performance and monitor usage patterns. The collected data remains anonymous and contains no sensitive information, protecting user privacy during the app's development and optimization process.
User interactions with ChatGPT within the Sidekick app are designed to be ephemeral and are not stored, either locally or on remote servers. This ensures that both prompts and completions remain private between the user and the OpenAI API.
The app stores sensitive information in the form of OpenAI API keys, which are securely maintained locally in the user's keychain. These keys never leave the device, providing an additional layer of security against unauthorized access or data breaches.
For analytics and crash reporting, Sidekick employs Firebase, which collects anonymous data to enhance the app's performance and monitor usage patterns. The collected information remains anonymous and contains no sensitive elements, maintaining user privacy throughout the app development and optimization process.
Based on the documentation provided, Sidekick's cost structure works as follows: the app charges $0.002 per 1,000 tokens, with no additional hidden fees. This means that every 1,000 characters (including spaces and punctuation) processed by the OpenAI API incurs a cost of two cents.
As an example, the developer mentions that 2,500 pages of text would cost $2 to process. To break this down further:
A single page of text typically contains around 500 words
2,500 pages therefore equates to 1,250,000 words
At 1,000 tokens per 1,000 words (considering punctuation and formatting), this would be 1,250,000 tokens
At $0.002 per 1,000 tokens, this would indeed total $2
The pricing model is designed to be straightforward and transparent, with the developer explicitly stating that they do not profit from app usage. All costs are directly tied to OpenAI's token pricing structure, giving users clear visibility into their expenses.
The app's performance data is collected through Firebase, which gathers anonymous usage information to improve Sidekick's functionality and monitor app performance. The collected data remains completely anonymous and contains no personally identifiable information or sensitive details.
The app stores sensitive information in the form of OpenAI API keys, which are securely maintained locally in the user's keychain. These keys never leave the device, ensuring that the app developer has no access to or retention of any authentication credentials.
The Sidekick developer explicitly states they do not profit from app usage, with all costs directly tied to OpenAI's token pricing structure. This transparent pricing model gives users clear visibility into their expenses, with no additional hidden fees.