Project Overview: I'm seeking an experienced developer or team who can help design and implement a system to allow users to leverage their existing AI subscriptions (e.g., OpenAI Plus, Anthropic, Google Gemini) for accessing language models through our custom hardware devices. The goal is to create a seamless experience where users can interact with AI through our hardware devices that are local on their network and rather than using our Api key to send/receive data, they alternatively use their own subscription credentials to access the llm models. We are agnostic on which llm to access. But it seems duplicitous to have users pay a subscription fee and then have to pay once again, to access the llm using our proprietary hardware. We can't figure this out, but maybe there is someone out there that could..!! is that you? .
Key Problem to Solve: Users already pay for subscriptions like OpenAI Plus to access LLMs. However, when they use our devices, we currently need to use our own API credentials to query these models, which results in extra, redundant costs. We want to eliminate this by having users query their personal accounts via our infrastructure, thus reducing fees on both sides. Since OpenAI at this time does not support Oauth, we are not sure how to manage this if it can be managed. It surely seems as though it should be able to be...
Scope of Work:
Develop a user portal where individuals can register and securely enter their subscription credentials for AI services (OpenAI, Anthropic, Gemini, etc.).
Enable seamless integration of these user accounts so that when they interact with our devices, the LLM queries are made via their personal subscriptions, rather than our own API keys.
Ensure secure handling and storage of user credentials, including encryption and privacy safeguards.
Implement token-based or OAuth-style authentication workflows, or any workaround that allows for secure, delegated access using users' subscription models.
Build an integration layer between our hardware devices and these user-provided credentials, allowing the device interactions to route queries securely and correctly.
Key Requirements:
Proven experience with OAuth 2.0, API authentication, or other authorization systems.
Knowledge of LLM API services like OpenAI, Anthropic, Google Gemini, etc., and their subscription models.
Experience with secure credential storage, encryption, and implementing privacy safeguards.
Understanding of user portal creation and authentication workflows that minimize friction for users.
Ability to work around or innovate on existing LLM access limitations to achieve the goal of using user-specific credentials instead of our API keys.
Knowledge of server-client interaction with custom hardware devices is a plus.
Ideal Candidate:
Creative problem solver who can think outside the box, given the current limitations of LLM subscription models.
Experienced developer familiar with authorization, encryption, and AI integration.
Capable of discussing and iterating possible solutions to explore the feasibility of this concept.
Additional Notes: This project is aimed at developing a functional prototype that can scale to thousands of users. If you have experience working on similar challenges or believe there is a workaround to achieve this type of delegated access, we are eager to hear from you. The current market does not support direct OAuth-like delegated access for these LLM subscriptions, so we are looking for someone who can think creatively and develop a possible solution.
Please include in your proposal:
Any relevant experience related to OAuth, API security, or LLM integration.
Thoughts on potential approaches for handling user-owned subscriptions in this manner.
Estimate of the time required to develop an initial proof of concept and potential pitfalls you foresee.
We are open to discussing innovative solutions and exploring what is technically possible. Let’s work together to push the boundaries and create something truly unique.