Leading LLMs, One API
Access DeepSeek, Qwen, GLM, Kimi, and other top models through a single OpenAI-compatible API — built for teams who need reliable routing, compliance, and unified billing.
OpenAI-compatible API · Multi-tenant · Pay-as-you-go
The Unified Gatewayfor EnterpriseLLMs
One Base URL. One auth model. One model catalog.Access DeepSeek, Qwen, GLM, Claude, and every major LLM through a single OpenAI-compatible endpoint.
One Base URL. One auth model. One model catalog.
Works with leading model providers
BUILT FOR PRODUCTION AI TEAMS
One gateway. Every model.
Leading LLMs, multi-tenant isolation, budget guardrails, and usage visibility — all behind one OpenAI-compatible API.
Multi-Tenant & Project Isolation
Organize work by organization and project with scoped API keys, isolated quotas, and clear ownership — ideal for platform teams serving many internal groups.
Budgets, Quotas & Alerts
Set budgets and usage limits per tenant or project. Get proactive alerts before overspend and enforce throttling or cutoffs to keep costs under control.
Usage Analytics & Dashboards
Visualize token usage, cost trends, and model breakdowns over time. Finance and engineering teams share one source of truth for AI spend.
REAL MODELS. REAL SERVICE.
Not a hollow gateway.
Callable models, metered usage, and traceable routing — every request is real.
- 01
Real model catalog
Models come from live provider integrations. GET /v1/models lists IDs you can actually call — not a static showcase.
- 02
Metered by usage
Every chat completion records token usage with quota and balance pre-checks. Console and billing stay aligned.
- 03
Smart model routing
Routes to available services by model id. X-SNX-* response headers help you diagnose issues.
// Call any model via OpenAI-compatible API
const client = new OpenAI({
baseURL: "https://api.supanexus.ai/v1",
apiKey: process.env.SUPANEXUS_KEY
});
const res = await client.chat.completions.create({
model: "deepseek-v3",
messages: [{ role: "user", content: "Hello" }]
});
PRIVACY
Zero retention of conversation content
Questions you send and answers the model returns are not kept on SupaNexus's side. We record only what billing and usage reports need, and forward conversations in real time to the model vendor you choose.
We don't save conversation text
Your questions and the model's replies are not stored in our systems. SupaNexus does not use your conversations for training, profiling, or secondary analysis.
Billing info only
We keep only what invoices and usage reports need — which model was used, how much was consumed, and when — not the words in your conversation.
Forward as-is, no tampering
Conversations go straight to your chosen model vendor in real time. SupaNexus does not cache, rewrite, or slip in extra content.
Credentials kept secure
Keys needed to reach model vendors are stored encrypted and used only to forward requests. They never appear in plain text in the console or logs.
PAY AS YOU GO
Marketplace pricing. No surprises.
No fixed bundles or hidden channel fees. Each model's Input / Output token rates are shown live in the model marketplace — pay only for what you use.
Live rates in the marketplace
Sign in to the console and open Models to browse listed models with Input / Output pricing before you call.
Input and Output priced separately
Billing follows token usage; input and output rates often differ and align with GET /v1/models metadata.
Usage and billing from one pipeline
Usage analytics, account balance, and settlement share the same metering chain — auditable in the console.
PAY AS YOU GO
Token-based pricing
Pay per token. No subscription required.
deepseek-v3
Input / 1M $0.27Output / 1M $1.10
qwen-max
Input / 1M $1.60Output / 1M $6.40
glm-4-plus
Input / 1M $0.50Output / 1M $0.50
kimi-k1.5
Input / 1M $0.60Output / 1M $2.40
gpt-4o
Input / 1M $2.50Output / 1M $10.00
claude-3-5-sonnet
Input / 1M $3.00Output / 1M $15.00
Illustrative prices only. Actual rates are shown in the live model marketplace.
GET STARTED
Simple 3-step setup.
From signup to your first API call in minutes.
1
Create account
Register in the developer console and set up your organization.
2
Get API key
Create a project-scoped API key with the permissions you need.
3
Call models
Point your client at the SupaNexus OpenAI-compatible endpoint and start inferencing.
FAQ
Common questions.
Everything you need to know about SupaNexus, billing, and the API.
SupaNexus adds multi-tenant isolation, unified billing, budget guardrails, and routing across many providers — all behind one OpenAI-compatible API.
Each organization is its own tenant. Projects within an org have separate API keys and quota pools that never share balances.
In the developer console, open your project and navigate to API Keys. Keys can be scoped with per-key rate limits and expiry dates.
Caps cascade: a request is blocked if any of the key, project, or org budgets is exhausted. The most restrictive cap wins.
Alerts notify you (email / webhook) when spend crosses a threshold but let traffic continue. Hard caps block new requests once the limit is hit.
Open the console dashboard, select a time range and model filter, then go to Budgets to set thresholds and notification channels.
Build in a weekend. Scale to millions.
One API key. Every major LLM. Make your first call in minutes.