Managed API BYOK

You know what your app does.
We route it.

One endpoint. Declare your task type. Oriora picks the best available model — scored across quality, cost, and latency, within the vendor preferences and priority you set — and gets out of the way.

Zero prompt storage — at every layer of the chain, by architecture

Your key, your call

Two ways to use it.

Use Oriora's routing intelligence whichever way fits your stack. Hand us your vendor key and we run the call — or keep your key entirely on your side and we just tell you which model to use. Same selection brain either way.

Server-side execution

You give Oriora your vendor key. We route to the best model, handle caching, fallback, and circuit breakers, and return the output. One request — we do the rest.

Orchestration fee on the call.

Client-side execution

You keep your vendor key on your own infrastructure. Ask us which model fits your task; we return a ranked recommendation. You make the call yourself. We never see your key, your prompt, or your output.

Flat fee per recommendation.

Both run on the same scoring brain — quality, cost, and latency across every supported vendor, within the preferences you set. The only difference is whether the AI call runs on our side or yours.

How it's different

Declare, don't classify.

Most routing tools analyse your prompt in real time using an ML classifier to guess what kind of task it is. We don't do that.

The classifier approach

  • Reads your prompt

    An ML model inspects every request to infer its task type. Your data is part of the classification pipeline.

  • Black box decisions

    You can't explain why a model was picked — the classifier decides and you trust it.

  • Added latency

    Every request pays a classification overhead before routing even begins.

  • Built for ambiguity

    Useful when task type changes per message. But most apps already know what they do.

The Oriora approach

  • You declare the task

    Your app is a code reviewer, or a support bot, or a research tool. It always will be. You pass taskType — we never need to read your prompt.

  • Fully auditable

    Every routing decision is explainable. Model X was picked because it scored highest on coding benchmarks at your tier.

  • Zero classification overhead

    No ML inference before routing. Your request goes straight to the best model.

  • Built for apps with a purpose

    If your app knows what it does, declare it. That's the only requirement.

Privacy

No prompt at rest. Anywhere.

Not a policy. Not a contractual add-on. The architecture itself has nowhere for your prompt to land.

Managed selection

Logs task type, token count, and cost per request. That's it. Prompt content is never written to disk or any database.

Model gateway

Forwards your request to the model provider. Configured for zero prompt retention — nothing stored in transit.

Model provider

Runs inference and returns a response. Same as calling them directly — the routing layer adds no extra data surface.

Relevant for any privacy-conscious productwhere prompt content shouldn't pass through a third-party logging layer. Some routing tools store prompts by default and charge extra for zero-data-retention. We don't store them at all.

Smart model selection, prompt cache pass-through, fallback, and circuit breakers across every supported vendor — full mechanics and current models on the pricing page.

How it works + supported models →

Up in three steps.

01

Create an account

Sign up at orioralabs.com. Connect your vendor API keys — Oriora charges only the orchestration fee.

02

Generate a key

Inside your account, generate an sk_oriora_... key. Yours in seconds.

03

Pass your task type

One POST request with taskType declared. Oriora handles everything from there.

Generate your API key →

Opens Settings → Oriora API keys (sign in required).

Developer reference.

Two endpoints, both authenticated with your sk_oriora_… key. Plain HTTP — no SDK required.

Client-side — get a model recommendation

Oriora returns the best-fit model; you make the call with your own vendor key. Your key never touches us. Flat $0.001 per recommendation.

curl https://api.orioralabs.com/api/select \
  -H "Authorization: Bearer sk_oriora_..." \
  -H "Content-Type: application/json" \
  -d '{"task_type": "coding", "models": ["openai/gpt-5", "anthropic/claude-opus-4-7"]}'

# → { "model": "openai/gpt-5", "alternatives": [...], "task_type": "coding" }
# "models" is optional — omit it and Oriora ranks across the full catalogue.

Discover task types (no auth): GET /api/select/task-types

Server-side — routed completion (OpenAI-compatible)

Point your existing OpenAI SDK at Oriora. We route to the best model and run it on your own vendor key (BYOK) — your vendor bills the tokens; Oriora charges only the platform & routing fees.

curl https://api.orioralabs.com/v1/chat/completions \
  -H "Authorization: Bearer sk_oriora_..." \
  -H "x-oriora-app: my-app" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "oriora-auto",
    "messages": [{"role": "user", "content": "Review this PR diff..."}]
  }'

# Returns an OpenAI-shaped chat.completion.
# model:"oriora-auto" lets Oriora route; x-oriora-app is the label you
# enabled BYOK for under Settings -> Oriora API keys / Provider keys.

Ready to route?

Get a key in seconds. First request in under five minutes.