Status: [UNTESTED] Endpoints found: http://localhost:8000/v1/chat/completions http://localhost:8000/api/v1/admin/keys http://localhost:8000/api/v1/admin/keys/ http://localhost:8000/api/v1/telemetry/stream Source: https://github.com/mylifeastanmay-hub/aegis-gateway # AegisGateway 🛡️⚡ [](https://github.com/mylifeastanmay-hub/aegis-gateway/releases/tag/v1.0.0) [](https://github.com/mylifeastanmay-hub/aegis-gateway/actions/workflows/ci.yml) [](https://fastapi.tiangolo.com) [](https://www.python.org) [](https://www.docker.com) [](https://redis.io) [](https://docs.pytest.org) [](LICENSE) > **Intelligent Zero-Trust LLM Gateway & Security Proxy** built with Python 3.11, FastAPI, and High-Performance Async HTTP Streaming (`httpx`). --- ## 🚀 Executive Summary Enterprise AI deployments face three critical operational threats: 1. **PII & Secret Compliance Leaks**: Credit card numbers, API keys, JWTs, and employee emails exposed in LLM prompts violate GDPR, HIPAA, and PCI-DSS compliance. 2. **Uncontrolled API Spend**: Identical, repetitive LLM prompts consume millions of redundant tokens, causing unexpected cloud billing spikes. 3. **Single-Provider API Outages**: Sudden 5xx errors or latency spikes from upstream providers crash customer-facing applications. **AegisGateway** acts as a unified zero-trust proxy layer between your application and upstream LLM providers (OpenAI, Groq, Anthropic, or local models). It delivers **sub-10ms reversible PII redaction**, **sub-5ms semantic caching**, **distributed token