Qorinix QX / PUBLIC
DE Chat Start building
QX-FASTNET INITIALIZING
UPDATED --:--:-- UTC
DOC / PRODUCT PUBLIC WEBSITE SPEED-FIRST INFERENCE REGION / LONDON-EDGE

High-Speed AI Inference Cloud

The High-Speed AI Inference Cloud

AI inference for interactive applications. Try Qorinix Chat and Arena free, or build on the OpenAI-compatible developer API with prepaid credit.

Start Building

From a first prompt to a first API call.

FREE ACCOUNT / API KEYS / DOCS
01

Create a free account

No card needed. Chat and Speed Arena work straight away, your conversations are saved to the account, and you can move to Go, Plus or Pro when you need higher daily limits.

02

Compare before you build

Prototype prompts in Chat, compare models side by side, and inspect response timing in the Speed Arena.

03

Call the API

Create a key, add prepaid credit from US$10, and point an OpenAI-compatible client at https://qorinix.com/v1. You pay only for the tokens you use. Text only for now: no tool calling, no images.

Speed Arena

Performance targets that map to real-time product pain.

Qorinix is designed for applications where waiting on inference breaks the experience. Metrics below are public product targets from the report and should be read as engineering targets, not historical production claims.

Metric Reference A Reference B Qorinix Target

Speed-Sensitive Markets

Built for applications where latency is product quality.

Map your workload

Explore assistants, support, gaming and translation workloads. Read inference use cases.

Runtime Architecture

One API over routing, cache, kernels, and execution.

Qorinix enters through a low-friction completions API, then compounds speed through routing, semantic cache, workload optimization, and deployment intelligence.

    QX RUNTIME FLOORPLAN LOAD / SYNTHETIC
    Speed path Cache path Execution tile

    Six-Layer Compound Moat

    Full-stack inference optimization, released progressively.

    PUBLIC TECHNICAL SUMMARY
    Layer Core technology Target metric Stage

    Product Roadmap

    From Speed Arena to production API and deeper infrastructure ownership.

    Public roadmap language is product-focused: launch the prototype, prove the API, expand developer adoption, and support enterprise deployment.

    Pricing

    Membership for Chat and Speed Arena. Prepaid credit for the API.

    MONTHLY MEMBERSHIP + PREPAID API CREDIT

    Membership: Free US$0, Go US$8, Plus US$20 and Pro US$50 per month. Developer API: from US$0.25 per 1M input tokens, with prepaid credit from US$10. View full pricing.

    Membership is billed monthly in US dollars and can be cancelled at any time. Daily limits reset at midnight UTC. Paid plans are subject to a monthly fair-use allowance: if it is used up, Free limits apply until the next month. Guests get 10 chat runs and 10 Arena races a day with no account, and a free account needs no card. The Developer API is separate from membership: membership does not include API credit. Compare every plan.

    Developer API

    OpenAI-compatible API, paid per token.

    Point an OpenAI-compatible client at https://qorinix.com/v1. Create a key in the developer console, add prepaid credit from US$10 (valid 12 months) and pay only for the tokens you use. Text only for now: no tool calling, no images. Membership does not include API credit. Read the API documentation.

    api / v1.chat.completions OPENAI-COMPATIBLE
    // Base URL
    https://qorinix.com/v1
    
    // Endpoints
    POST /v1/chat/completions
    GET  /v1/models
    
    // Price per 1M tokens (input / output)
    "qorinix-3.1-instant":   US$0.25 / US$1.00
    "qorinix-3.2-reasoning": US$1.00 / US$3.00
    
    // Request
    Authorization: Bearer YOUR_API_KEY
    {
      "model": "qorinix-3.1-instant",
      "messages": [{ "role": "user", "content": "Hello" }]
    }
    PREPAID CREDIT FROM US$10 / PAY PER TOKEN / TEXT ONLY QX

    Contact

    Bring a latency-sensitive workload.

    Share your traffic profile, response target, and model needs. Qorinix routes public inquiries through the main company contact channel.

    Company
    Qorinix Limited
    Address
    71-75 Shelton Street, Covent Garden, London, United Kingdom WC2H 9JQ

    Ready

    FAQ

    Questions serious teams ask before moving latency-critical AI into production.

    Discuss your workload
    Is the public Qorinix API already open?

    Yes. The Developer API is OpenAI-compatible at https://qorinix.com/v1 and is paid for from prepaid credit, from US$10. Create keys at /account/api and read the documentation at /docs/api. Text only for now: no tool calling, no images.

    What does the Arena prove?

    The Arena is a live, repeatable comparison of TTFT, total latency, throughput, and output size across Qorinix-managed lanes and neutral reference lanes. It is built to show product-level responsiveness without exposing routing internals.

    Can Qorinix work with an existing chat-completions client?

    Yes. Point the client at https://qorinix.com/v1 and call POST /v1/chat/completions with the model qorinix-3.1-instant or qorinix-3.2-reasoning. GET /v1/models lists the models. Text only for now: no tool calling, no images.

    How are model credentials and routing policies protected?

    Credentials, exact model IDs, route weights, and resilience policy stay server-side. Public pages use neutral reference names only, and client-side code does not contain private API keys or commercial routing details.

    What kind of workloads fit Qorinix best?

    Qorinix is strongest where response time is part of product quality: voice agents, customer support automation, game NPCs, trading alerts, real-time translation, recommendation systems, and high-volume chatbot traffic.

    How should a team start?

    Start with Chat or the Arena: guests get 10 chat runs and 10 Arena races a day with no account. A free account needs no card, and the Go, Plus and Pro plans raise the daily limits. To build, create an API key, add prepaid credit and call the API; membership and API credit are separate.