Zero data retention LLM API

One API key.
Every model.
No retained data.

Route your requests to a curated list of zero data retention providers through one OpenAI-compatible endpoint. Your prompt is processed and dropped. It is not logged, not stored, and not trained on.

Get API key Read the docs

OpenAI-compatible No subscription Pay per token

POST api.zeroretentionapi.ai/v1/chat/completions

# Any model from the catalog works here.
curl https://api.zeroretentionapi.ai/v1/chat/completions \
  -H "Authorization: Bearer $ZERO_RETENTION_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "provider/model-name",
    "messages": [
      { "role": "user", "content": "Hello" }
    ]
  }'

import os
from openai import OpenAI

client = OpenAI(
    base_url="https://api.zeroretentionapi.ai/v1",
    api_key=os.environ["ZERO_RETENTION_API_KEY"],
)

# Swap the model name. The retention policy does not change.
response = client.chat.completions.create(
    model="provider/model-name",
    messages=[{"role": "user", "content": "Hello"}],
)

print(response.choices[0].message.content)

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api.zeroretentionapi.ai/v1",
  apiKey: process.env.ZERO_RETENTION_API_KEY,
});

const response = await client.chat.completions.create({
  model: "provider/model-name",
  messages: [{ role: "user", content: "Hello" }],
});

console.log(response.choices[0].message.content);

retained: 0 bytes logged: no trained on: no

Curated list of zero data retention providers

0 bytes kept after the response is sent

100% of the catalog sits under a no-retention agreement

One endpoint your SDK code stays the same

No subscription prepaid credits, pay per token

The retention question

Where your prompt goes, and what stays behind.

A retention policy is a promise about the inside of someone else's data centre. You cannot see it. So the useful question is not “do they keep it” but “can I know what happens”.

Direct to a provider Unverified

Retention: Unknown
Prompt logging: Unknown
Training on your data: Unknown
Your proof: A policy page

Policy differs per provider, per model, and sometimes per endpoint. Reading it is all you can do.

Through ZeroRetentionAPI Known

Retention: None, under agreement
Prompt logging: None
Training on your data: Never
Your proof: The agreement, per provider

Every provider in the catalog is covered before it takes a single request. A provider that will not agree does not enter the catalog.

We do not claim that other providers keep your data. We claim that you cannot know. This API removes the guessing.

What you get

Built for teams that cannot keep the data.

OpenAI-compatible

Change the base URL and the key. Your SDK, your framework, and your prompts stay as they are.

Zero retention routing

A request only reaches a provider that has agreed to keep nothing. There is no setting to forget.

One key, many models

Small, fast, and large models from different providers, behind a single credential and one bill.

Prepaid, no subscription

Add credits when you need them. Pay per token. No monthly fee and no seat count.

Our three promises

What we never do.

We never store your prompt. The request lives in memory for the length of the call. Then it is gone.
We never train on your data. We cannot train on data that we do not keep.
We never sell or share it. No data broker, no reseller, no advertising use. Ever.

No compliance badge appears on this page, because we have not earned one yet. Ask us anything about data handling at hello@zeroretentionapi.ai.

Pricing

Prepaid credits.
No subscription.

Four steps from zero to your first response. You pay for the tokens you use, at the price of the model you call.

STEP 01
Sign up

Create an account with a work email.
STEP 02
Add credits

Top up the balance. There is no free tier and no card on file.
STEP 03
Create a key

Name the key, copy it once, keep it in your secret store.
STEP 04
Call the API

Send the request to the endpoint. Your balance drops by the token cost.

Get API key Volume pricing

Enterprise solution

For teams that need it
in writing.

Volume traffic, a written no-retention commitment, and a person who answers. Tell us what your legal and security teams need to see.

Volume pricing on committed spend
A written no-retention commitment, per provider
Invoice billing instead of prepaid credits
A direct line to the engineers who run the routing

Questions

Straight answers.

What does “zero retention” mean here?

Your prompt travels to a provider, the answer comes back, and neither is written to storage. Nothing about your request stays on our side once the response is delivered.

Do you log my prompts?

No. A log is a copy, and a copy is retention. We keep request metadata for billing only: the model, the token count, and the time. The prompt text and the answer are not in it.

Which providers do you route to?

A curated list. Each provider signs a no-retention agreement before it takes traffic. A provider that will not agree does not enter the catalog, whatever its public policy says. See Models for the current list.

Is it really OpenAI-compatible?

Yes. The request and response shapes follow the OpenAI chat completions schema, so the official SDKs and the common frameworks work after a base URL change. See Docs for the exact endpoint.

How does billing work?

Prepaid credits. Sign up, add credits, create a key, then call the API. You pay per token at the price of the model you call. There is no subscription and no free tier. Enterprises can ask for invoice billing under Enterprise Solution.

What happens to my data if I stop using the service?

Nothing to delete. Your account holds billing records, not prompts. Close the account and those records go with it.