02 — Smart Router · 12+ Providers · Real-Time Scoring

The Right Model.
The Right Price.
Every Single Call.

Smart Routing Consensus Mode BYOK Cost Optimisation Compliance Filters

Every query scored across 12+ providers in real time — on cost, quality, latency, and compliance. You set the weights. We handle the selection. Stop overpaying. Stop thinking about which model to use.

// 12 providers · 38ms avg routing latency · 40–65% average cost reduction

AGENT ENDPOINT POST https://api.eightx.app/api/gateway/route · auth: Bearer <api_key> · modes: smart · manual · consensus full docs →
12+
LLM Providers
38ms
Avg Routing Latency
40–65%
Avg Cost Reduction
4
Routing Modes
0.5%
Min Routing Fee

Three Ways to Route.
One API to Learn.

Whether you want full autonomy, manual control, or consensus confidence — the same endpoint handles all three. You pass a flag. We do the rest.

Smart Routing
mode: "smart"
Tell us the task. We score every available provider on cost, quality, latency, and your compliance requirements — and route to the winner automatically. No model knowledge required. The playground shows you what we selected and why.
→ Best for: most use cases
POST api.eightx.app/api/gateway/route
docs →
🎯
Manual Selection
mode: "manual"
You specify the model. We route through it. Still get EightX's governance layer, audit trail, and Agent Passport attachment — just without the automatic selection. Good for workflows with specific model dependencies or contractual requirements.
→ Best for: specific model requirements
POST api.eightx.app/api/gateway/route
docs →
🔀
Consensus Mode
mode: "consensus"
Submit to multiple models simultaneously. Get each model's answer plus an agreement score (0–100) and a synthesised response. High agreement means trust it. Low agreement means review the individual answers. Built for irreversible or high-stakes decisions.
→ Best for: critical decisions, legal, medical
POST api.eightx.app/api/gateway/route
docs →

Scored on Four
Dimensions. Every Time.

The router doesn't just find the cheapest model. It weights quality, latency, cost, and your active compliance filters simultaneously — and selects the provider that maximises the outcome for your specific query.

💰
Cost
Per-token pricing across all providers, updated in real time. We always know the current rates.
Quality
Model benchmarks, task-type suitability, and live error rates weighted per query type.
Latency
Time-to-first-token and throughput per provider, measured continuously across routing calls.
⚖️
Compliance
GDPR, HIPAA, SOC 2 filters applied before scoring — non-compliant providers are excluded automatically.
See It Live in the Playground →
▶ Your Query
↓ submitted to
⚡ EightX Smart Router
↓ scoring 12 providers in real time
Claude 3.7 Sonnet★ 94 · $0.0008/1k
GPT-4o★ 91 · $0.0025/1k
Gemini 2.0 Flash★ 88 · $0.0003/1k
↓ winner selected — 38ms
✓ Optimal result returned · Passport attached

Already Have Enterprise
Contracts? We Sit On Top.

If your organisation already has enterprise deals with OpenAI, Anthropic, or Google — we don't replace them. We make them better. Bring Your Own Keys and get EightX's routing intelligence, governance, and cost attribution on top of the rates you already negotiated.

  • 🔑
    Complete Key IsolationYour keys are encrypted at rest. We never log them, store them plaintext, or access them outside the API call window.
  • 📊
    Full Cost AttributionPer-agent, per-department, per-query spend analytics — across every provider you're using, in one dashboard.
  • ⚖️
    Governance EverywhereOne compliance layer across every model and every provider contract — HIPAA, GDPR, SOC 2, all applied uniformly.
  • 🏪
    Marketplace AccessYour agents still access EightX's marketplace for services and capabilities your enterprise contracts don't cover.
$40K+
Average annual savings for enterprise teams who add EightX routing on top of existing provider contracts.
Talk to Enterprise →
⚡ EightX Smart Router
Routing · Scoring · Governance · Attribution
↕ on top of your existing contracts
OpenAI
Enterprise
Anthropic
Enterprise
Google
Vertex AI
↕ keys isolated in
🔐 Encrypted Key Vault
Zero-access architecture · Agent-level isolation

// BYOK pricing:
Standard routing fee applies on calls we route — even over your own keys. The governance layer, attribution dashboard, and marketplace access are included. No additional licence fee.

12 Providers. One API.
Zero Lock-In.

The moment a better model ships, we add it. You call the same endpoint. The router does the rest.

🧠
OpenAI
GPT-4o · o3 · Mini
🤖
Anthropic
Claude 3.7 · Haiku
🔮
Google DeepMind
Gemini 2.0 · Flash
🌬️
Mistral AI
Mixtral 8x7B
🦙
Together AI
LLaMA 3.1 405B
🧬
Cohere
Command R+
Cerebras
Llama 3.1 70B
☁️
AWS Bedrock
Nova Pro · Titan
🔵
Azure AI
Phi-4 · Copilot
🤗
Hugging Face
Open Models
🔥
Replicate
Flux · Open Models
🚀
Fireworks AI
Fast Inference
+
More Coming
New providers added continuously. Request one.
Try It Free in the Playground →

CODE EXAMPLE

    
Full Docs →
💬 EightX Feedback Beta
EightX is in beta — help us improve! Share bugs, recommendations, or anything on your mind. 🙌
Feedback