llmcloud.ai
Platform · BYOK

Your keys. Our router.

Bring your own OpenAI, Anthropic, Google, xAI, Fireworks, Together or Groq credentials. Keep your pricing and residency — routing is free on the free tier.

pricing

Zero markup on tokens

The router forwards to your provider account. You're billed by the provider at your rate — no take-rate.

residency

Keep your contracts

Enterprise agreements, DPAs, zero-retention and regional pinning still apply. The router is transport.

credits

Use existing credits

Spend Anthropic, Azure and Google credits inside the router without duplicating budget.

hybrid

Hybrid billing

Mix BYOK and managed. Route auto:quality to your Claude account, auto:cost to pooled DeepSeek.

secure

Envelope encryption

Provider keys are sealed with per-org KMS, decrypted only at request time. Never logged.

rotate

Rotate anytime

Rotate upstream keys in the dashboard or API. The router picks up new keys on the next request.

Attach a provider key
POST /v1/provider-keys
{
  "provider": "anthropic",
  "credential": "sk-ant-...",
  "scope": { "keys": ["agent-prod-router"] },
  "region": "us-east"
}
Route through it
POST /v1/chat/completions
{
  "model": "anthropic/claude-4.5-sonnet",
  "messages": [...],
  "route": { "byok": true }
}
Do I still get analytics and audit logs on BYOK traffic?+

Yes — tokens, latency, cost, cache status and guardrail signals, same as managed.

What providers are supported?+

OpenAI, Anthropic, Google (Vertex + Gemini), xAI, Groq, Cerebras, SambaNova, Fireworks, Together, Deepinfra, Azure OpenAI. Custom OpenAI-compatible endpoints on request.

Can I run BYOK-only?+

Yes. The free tier gives you routing, caching, MCP, guardrails and analytics on top of your own accounts, with no markup.