Does Zima store my data?

No. Zima enforces zero data retention across all models. For open-source models, requests are processed inside hardware-encrypted enclaves using Intel TDX and NVIDIA Confidential Computing. For centralized models like GPT and Claude, requests are routed with zero logging and zero data retention.

Is Zima compatible with the OpenAI API?

Yes. Zima provides an OpenAI-compatible endpoint. Switch your API base URL and key — no other code changes needed. Works with any SDK or framework that supports the OpenAI API format.

How is Zima different from using OpenAI directly?

Zima adds hardware-level encryption and zero data retention to your AI API calls. With OpenAI directly, your prompts may be logged. With Zima, nothing is stored — and for open-source models, processing happens inside encrypted hardware enclaves.

What is confidential computing?

Confidential computing uses hardware security features like Intel TDX and NVIDIA Confidential Computing to encrypt data while it is being processed — not just at rest or in transit. Zima uses this to ensure your AI workloads remain private even from infrastructure operators.

Which models does Zima support?

Zima supports 100+ models including GPT-4, Claude, Gemini, Llama, Mistral, DeepSeek, Qwen, and more. Open-source models run on encrypted infrastructure with full hardware isolation. New models are added weekly.

Hardware encrypted inference.

No logs. No limits.

An encrypted AI gateway for every model. Drop-in OpenAI SDK replacement with zero data retention.

Create Key Documentation

Integration

Secure computation, end-to-end

Inference leading LLMs run inside verifiably secure runtimes, powered by Intel TDX and NVIDIA Confidential Computing architectures.

inference.ts

1import OpenAI from "openai"
2
3const zima = new OpenAI({
4  baseURL: "https://www.zima.chat/api/v1",
5  apiKey: process.env.ZIMA_KEY,
6})
7
8const response = await zima.chat.completions.create({
9  model: "qwen3-next-80B-a3B-instruct",
10  messages: [{ role: "user", content: "..." }],
11})
12
13// encrypted in hardware. zero data retained.

www.zima.chat

POST /api/v1/chat/completions

host www.zima.chat

authorization zima_sk_•••dk8f

content-type application/json

Sealed payload

{ "model": "qwen3-next-80B-a3B-instruct",
  "messages": [{ "role": "user", "content": "..." }] }

✓ 200 OK · 138ms

  "model": "qwen3-next-80B-a3B-instruct",
  "encrypted_response": true,
  "no_data_stored": true,
  "usage": { "prompt_tokens": 512, "completion_tokens": 256,
             "total_tokens": 768 },
  "billing": { "cost_usd": 0.001384 }

Zima Protected

Encrypted inference verified

Encrypted: true
Data Stored: None
Tokens: 768
Cost: $0.00138

All inference is encrypted in hardware. Zero data retained.

Hardware-grade protection for every inference request.

Infrastructure

Inference you can count on

Hardware Encryption

Silicon-level TEE enclaves protect every inference request. Data encrypted in memory, sealed from host infrastructure.

Zero Retention

No prompts, outputs, or metadata survive past completion. Memory cryptographically wiped after every request.

OpenAI Compatible

Drop-in replacement for the OpenAI SDK. Switch your base URL, keep your existing code and tooling.

100+ Models

OpenAI, Anthropic, Google, Meta, and Mistral through a single encrypted endpoint. One key, zero vendor sprawl.

#Architecture

Verified Inference Path

From SDK request to protected output, without storing prompts or responses.

01Authenticate

API key verified, team resolved, usage policy enforced before entering the secure network.

02Encrypt

Payload sealed inside a hardware-backed TDX enclave. Data encrypted in CPU/GPU memory.

03Route

Forwarded to any of 100+ models. Provider credentials isolated, never exposed to the host.

04Deliver

Response returned, usage metered. Memory wiped per request - nothing persists.

#Models

Model Pricing

Provider costs passed through at-rate. Zero markup.

Model	Provider	Input / 1M	Output / 1M
GPT-4o	OpenAI	$2.50	$10.00
Claude Sonnet 4	Anthropic	$3.00	$15.00
Gemini 2.5 Pro	Google	$1.25	$5.00
Llama 3.1 70B	Meta	$0.50	$0.75
Mistral Large	Mistral	$2.00	$6.00

Browse all 100+ models

Your prompts never leave the enclave

Intel TDX + NVIDIA confidential compute — data encrypted in memory, sealed from host infrastructure.

Even we can't read it.

1 / 5

“We evaluated six AI gateways. Zima was the only one where our security team signed off without a single objection.”

Sarah Chen

VP Engineering, Meridian Health

#FAQ

Frequently asked questions

Security, compatibility, and pricing.

No. Requests exist only for the duration of execution inside the enclave. No prompts, outputs, metadata, or logs are persisted. This is enforced by hardware, not policy.

Access is blocked immediately. If the system cannot cryptographically verify that the runtime environment is running approved, unmodified software, no data is processed.

Yes. Zima uses the chat/completions format. Point your existing OpenAI SDK at our base URL and it works — same request shape, same response shape, zero retention.

Every model in the catalog is labeled with its security tier. TEE-hosted models run entirely inside hardware enclaves. Proxied models are routed through encrypted transit but execute on the provider’s infrastructure.

Per-token, per-model, pay-as-you-go. No platform fees, no hidden tiers, no commitments. Pricing is displayed on every model card before you send a request.

Get Started

Start encrypting your inference today.

Point the OpenAI SDK at Zima, keep model choice and pricing visibility, and move sensitive traffic into attested hardware.

Create Key