UPDATE
AI GOVERNANCE MIDDLEWARE
For some reason, I found new interest in the project and started working on it again—yeah, I know this doesn’t count for the original submission—but I ended up building MVP 1.0 that actually works.
What's New?
I’ve added support for five major providers:
- OpenAI
- Anthropic
- AWS Bedrock
- Google Vertex AI
- Gemini
It’s fully expanded and testable now. Users can bring their own API keys and plug them in to test requests across multiple LLM providers.
Key Features
Functional and live:
https://5viugclv4yi2jrgt5dlcjxe6se0ntoto.lambda-url.us-east-1.on.aws/Focused on security and privacy:
No user dashboard, no account linking. Users can create their own API keys just by hitting the endpoint. Everything is included in the docs.Stateless by design:
Except for storing API keys, which are encrypted and stored in DynamoDB (since AWS Lambda doesn’t support writes).Minimalist, privacy-first architecture:
The whole system is designed without tracking, dashboards, or analytics—just clean middleware to help you test, evaluate, or build with AI providers under your own terms.
More improvements coming soon.
Note:
The [link I submitted earlier for the hackathon] and the GitHub repo remain unchanged. This new work was done recently and is not part of the official submission.
Log in or sign up for Devpost to join the conversation.