AI Gateway Platform for Multi-Provider Apps

A production-shaped gateway that routes requests across OpenAI, Anthropic, Gemini, and local mock providers with API keys, rate limits, prompt logs, cost tracking, and policy-aware orchestration.

0

Total Runs

0

Sessions

0

Telemetry Events

Core Capabilities

Runtime Status

Default Router

API Keys

Prompt Logs

Cost Today

Run Timeout

Max Plan Steps

Telemetry Retention

First Token p95

Canary Traffic

Human Review Mode

System Architecture

Gateway API

REST + SSE for sessions, runs, streaming tokens, feedback, and provider-specific requests.

Provider Router

OpenAI, Anthropic, Gemini, and mock providers sit behind one project-safe adapter contract.

Policy + Guardrails

Hard enforcement, prompt-injection neutralization, scope checks, and output redaction.

Telemetry + Eval

Run events, prompt hashes/previews, cost records, and offline evaluation for regression control.

Human Review Layer

Escalation queue and decision APIs let reviewers approve, reject, or request changes.

Recruiter Snapshot

This project demonstrates modern AI gateway engineering: model-provider abstraction, secure API access, rate limiting, cost awareness, prompt observability, deployment readiness, and measurable quality controls.

Open Live Health Endpoint Open Ops Dashboard