Features
Everything you need to ship AI in production
One endpoint, every model, none of the plumbing. Mappace is the API gateway that sits between your code and every LLM provider.
One endpoint for every model
A single OpenAI-compatible base URL fronts Claude, GPT, Gemini, Llama, Qwen, and your self-hosted endpoints. Swap models by changing one field — no new SDKs, no vendor lock-in.
Streaming & tool use
Native Server-Sent Events streaming and tool/function calling through the standard chat completions shape. Your existing OpenAI code runs unchanged — just repoint the base URL.
Automatic failover
If an upstream provider errors or rate-limits, the gateway retries the next healthy route so your users never see a hard failure. 502/504s are handled server-side.
Per-model context limits
Each model carries its own real context window. The gateway enforces it, auto-compacts long conversations server-side, and shows a live context meter so generation never silently truncates.
Scoped API keys & BYOK
Issue per-project keys you can revoke instantly. Bring your own provider keys and route spend through your own accounts while keeping one unified interface.
Consolidated billing
One balance, no markup on token usage. You pay exactly what the underlying model costs — top up any amount and only spend on the tokens you send.
Live model catalog
Browse every available model with context size, pricing, and provider in one place. The catalog updates as new models are added.
FAQ
Questions, answered
Do I need to change my OpenAI SDK code?+
No. Mappace is wire-compatible with the OpenAI chat completions API. Set the base URL to https://mappace.com/v1 and keep your existing SDK, requests, and streaming code.
What happens if a provider goes down?+
The gateway detects upstream errors and rate limits and fails over to the next healthy route automatically, so your requests keep succeeding without code changes on your side.
Can I use my own provider API keys?+
Yes. Bring-your-own-keys (BYOK) lets you route requests through your own provider accounts while keeping a single unified endpoint and dashboard.
Is there a markup on token usage?+
No. You pay exactly the underlying model price for each request. Top-up your balance once and spend only on the tokens you use — no markups, no recurring fees.
Ready to point your SDK at one endpoint?
Get started