Introduction
Welcome to ModelHive — the intelligent LLM routing, caching, and management platform.
What is ModelHive?
ModelHive is a unified AI gateway that sits between your applications and LLM providers (OpenAI, Anthropic, Google Gemini, Mistral, Groq, Azure, and more). It provides a single OpenAI-compatible API endpoint that intelligently routes, caches, compresses, and secures every request.
Key Features
| Feature | Description |
|---|---|
| HiveCache | Dual-embedding semantic cache that returns cached responses for semantically similar requests — DIRECT, REUSE, TWEAK, and MISS bands |
| HiveState | Adaptive state compression that reduces long conversation histories by up to 71% while preserving executable workflow state |
| HiveRoute | Zero-cost intelligent model routing that automatically selects the cheapest adequate model based on task difficulty |
| Security Guardrails | Real-time protection against prompt injection, toxicity, PII leakage, and abuse — configurable per tenant |
| Multi-Provider Routing | Route to 100+ models across OpenAI, Anthropic, Google, Mistral, Groq, Azure OpenAI, and more via a single API |
Architecture Overview
Key Capabilities
| Feature | Description |
|---|---|
| OpenAI-Compatible API | Drop-in replacement — use the standard OpenAI SDK with your ModelHive API key |
| Multi-Tenant | Isolate teams with separate budgets, keys, models, and guardrail configs |
| Real-Time Budget Control | Per-key budgets with wallet system, auto-recharge, and alerts |
| HiveCache | Dual-embedding semantic cache with DIRECT / REUSE / TWEAK / MISS routing bands |
| HiveState | Adaptive state compression — up to 71% token reduction for long conversations |
| HiveRoute | Zero-cost intelligent routing — trivial/standard/complex difficulty classification |
| Guardrails | Prompt injection detection, toxicity filtering, PII anonymization, secrets scanning |
| Compression | Automatic conversation history compression for token savings |
| Analytics | Comprehensive dashboard with spend tracking, cache savings, security events |
| Subscription Tiers | Free, Premium, Enterprise — each with different cache discount rates |
Quick Links
- Quick Start Guide — Get your first API call running in 2 minutes
- HiveCache Deep Dive — Understand our semantic caching technology
- API Reference — Complete endpoint documentation
- Security Guardrails — Configure protection for your LLM calls