Kynetra Conflux — One key. Every model. Launching soon.

Quickstart

Drop-in for the OpenAI SDK

Conflux speaks the OpenAI chat completions format. Point your existing OpenAI client at Conflux's base URL and swap in a Conflux key — nothing else in your code has to change.

Conflux isn't publicly available yet — the snippet below is a preview of the interface it will expose at launch, not a live endpoint you can call today.

from openai import OpenAI

client = OpenAI(
    base_url="https://conflux.kynetra.dev/v1",
    api_key="qai_...",
)

response = client.chat.completions.create(
    model="anthropic/claude-sonnet-4.5",
    messages=[
        {"role": "user", "content": "Say hello from Conflux."}
    ],
    stream=True,
)

for chunk in response:
    delta = chunk.choices[0].delta.content
    if delta:
        print(delta, end="", flush=True)

curl https://conflux.kynetra.dev/v1/chat/completions \
  -H "Authorization: Bearer qai_..." \
  -H "Content-Type: application/json" \
  -d '{
    "model": "anthropic/claude-sonnet-4.5",
    "messages": [
      {"role": "user", "content": "Say hello from Conflux."}
    ]
  }'

Keys will be issued from the QuantumOS admin console (prefixed qai_) once Conflux launches. At that point, conflux.kynetra.dev will be served by the QuantumOS admin host — the same /v1/chat/completions and /v1/models endpoints shown above. Set stream: true for server-sent-event streaming, as shown above; omit it for a single non-streaming JSON response.

Capabilities

What's actually in Phase 0

No roadmap items below — this is what the API is built to do, ahead of launch.

OpenAI-compatible

Standard /v1/chat/completions request/response shape, so existing OpenAI SDKs and tooling work with a base-URL and key swap.

Streaming and non-streaming

Set stream: true for server-sent-event token deltas, or omit it for a single JSON completion.

Unified credit billing

One credit balance across every model you call — no separate accounts or invoices per provider.

Per-key rate limits

Every key gets its own token-bucket limit, reported via standard RateLimit-* response headers.

Multi-provider routing

Requests route through OpenRouter to a wide model catalog, with bring-your-own-key (BYOK) support for your own provider accounts, plus support for local/self-hosted models via Ollama.

Usage dashboard

Per-key call counts, token usage, and cost show up in the QuantumOS admin console as they happen.

Models

Served live from `/v1/models`

The catalog isn't a static page — call GET /v1/models and you get the current list of active models with per-token pricing, in the OpenAI list format. The rows below are illustrative examples of the shape, not a live quote — check the endpoint for current prices.

Model id	Owned by	Context	Pricing (illustrative)
anthropic/claude-sonnet-4.5	anthropic	200k	example only
openai/gpt-4o-mini	openai	128k	example only
meta-llama/llama-3.1-70b	meta-llama	128k	example only
…	…	…	…

Illustrative rows only — exact model ids, context lengths, and prices are set by the live catalog at /v1/models, not by this page.

Pricing

Credit-based, pay for what you use

Pay per use. Every call debits credits based on actual input/output tokens at that model's catalog rate — no seat licenses, no minimum commitment.
Margin over upstream cost. Conflux prices in a margin above the underlying provider token cost; the exact rate per model is what /v1/models publishes.
Credit packs in the console. Top up your balance from the QuantumOS admin console; new keys start with a small starter balance for testing.
Finalized at GA. We're pre-launch — pricing mechanics (credits, margin, packs) are what's described here, but exact numbers are still being finalized ahead of general availability.

FAQ

Common questions

Does it really work with the OpenAI SDK?

Yes — point the OpenAI Python (or any OpenAI-compatible) SDK at https://conflux.kynetra.dev/v1 with a Conflux qai_ key as the API key. The request and response shapes for /v1/chat/completions follow the OpenAI format, including streaming via stream: true.

What happens when my credits hit zero?

Calls return an HTTP 402 Payment Required response (RFC 7807 application/problem+json body) until you top up. You won't be billed past your balance — the credit check runs before the model call goes out.

Is streaming supported?

Yes. Set stream: true in your request and Conflux returns server-sent-event chunks in the OpenAI chat.completion.chunk format, ending with a standard [DONE] frame.

Can I use my own provider API keys (BYOK)?

Yes — bring-your-own-key lets requests route through your own provider account instead of the shared OpenRouter pool for a given model, configured from the admin console.

How is my data handled?

Requests are logged for billing and usage purposes (token counts, cost, model, timestamps) and are visible in your own workspace's usage dashboard. We're not making broader privacy or compliance claims at this stage — ask if you need specifics for your use case.

One key. Every model.