MoYo is a production-grade AI API gateway. Fully compatible with OpenAI and Anthropic protocols — point your baseURL here and reach GPT-6 Astra, Claude Opus 5.5, Grok 4.7 and Gemini with load balancing and automatic failover built in.
Every call is scheduled across a pool of upstream channels by weight and live health. Knock a channel offline below — traffic reroutes in milliseconds and your app never notices.
An intelligent scheduling layer built for production traffic — protocol-perfect, observable end to end, and quiet when it matters.
Chat Completions, Responses and Anthropic Messages — streaming, function calling and multi-turn all pass through verbatim. Migrate the OpenAI SDK, Claude Code or LangChain by changing one line.
Degraded channels are bypassed in milliseconds. No request is dropped.
Isolated tokens, quota caps, per-key usage and cost analytics.
Full-path SSE passthrough keeps first-token latency close to calling upstream directly — long outputs stay smooth, end to end.
Request-level logs with status, latency and tokens — every call is queryable from the console.
Requests are relayed for protocol conversion only — nothing is stored or used for training. Keys are encrypted and revocable anytime.
Chat, reasoning, code and multimodal across four flagship families. Switch models by changing a single field — keys and code stay untouched. Click a card to copy its ID.
More models onboard continuously — upstream channels are handled for you.
Register, open the console and generate an API key. Your quota is visible at a glance.
Aim the SDK at the MoYo gateway. The wire protocol is identical to the official APIs.
Call any model by name. Streaming responses come straight back — hit Run to try it.
Anything that accepts a custom base URL works with MoYo — SDKs, agents, IDEs and chat clients. Two environment variables and you're done.
OPENAI_BASE_URL=https://v1.cdks.work/v1ANTHROPIC_BASE_URL=https://v1.cdks.workMOYO_API_KEY=sk-your-moyo-keyAt the protocol layer, it isn't. OpenAI /v1/chat/completions and /v1/responses, plus Anthropic-native /v1/messages, are passed through verbatim — same request shape, same response format, streaming included. Point baseURL at MoYo and you gain load balancing, failover and unified billing across upstream channels, without maintaining a key per vendor.
All four flagship families — GPT-6 / Claude 5.5 / Grok 4.7 / Gemini 3. Switching is a one-field change to model; your key and code stay untouched.
Gateway overhead stays under 80 ms. A multi-channel pool auto-selects the healthiest route with hot failover, targeting 99.9% availability. Streaming passes through end to end, so long outputs never drop.
Yes. Set ANTHROPIC_BASE_URL=https://v1.cdks.work and use your MoYo key as the auth token. Any tool that accepts a custom OpenAI or Anthropic base URL works the same way.
Requests are forwarded for protocol conversion only — nothing is stored, used for training, or retained for analysis. Keys are encrypted at rest and can be revoked or capped at any time.
No credit card required — register and go. One key, and multi-channel maintenance is over.