One key · seven providers

One key.
Every model.

OpenAI, Moonshot, MiniMax, DeepSeek, Grok, MiMo and GLM through one key. Keep your existing SDK, get one bill, and see what each request costs.

Get API key Read the docs

OpenAI & Anthropic compatible · Hash-only keys · Per-request usage
Invite-only during launch. An invite code is required to register.

7 providers · one key

One key · metered per token

Drag to rotate · Hover a beam

One endpoint

OpenAI & Anthropic compatible

Hash-only keys

Per-request usage records

Model access

Seven providers. One key.

Use OpenAI, Moonshot, MiniMax, DeepSeek, Grok, MiMo and GLM through one API key. Keep one Base URL and one bill; switch models without opening another provider account.

View the model catalog

OpenAI

Moonshot

MiniMax

DeepSeek

Grok

MiMo

GLM

Why SeedOne

Switch models without rebuilding your integration.

One key across providers

Use one credential and one Base URL across seven model families. Change the model name; your integration stays the same.

Keep the SDK you already use

Point the official OpenAI or Anthropic SDK at our compatible endpoint. A protocol mismatch is rejected explicitly, so it does not fail silently.

See what each request costs

Each request records the model, token counts, cost and status. Prompt and response bodies are not stored.

Only you can see your full key

Keys are stored as one-way hashes, visible to their owner and revocable in one click. Platform staff, exports and logs never contain the full value.

Give access without sharing your key

The creator owns the team wallet and sets per-key limits. When a member is removed, their team keys stop working.

A failed upstream does not have to fail your request

Upstreams are chosen by health and retried within a fixed budget. If every attempt fails, you get an explicit error with a request id.

Quickstart

From zero to a first token.

Create an API key

Sign in to the console, create a key, and scope it to the model families it may use. The full value remains visible to you, and only to you.

Point the SDK at one Base URL

Keep the official OpenAI or Anthropic SDK. Replace the Base URL and API key; the rest of your integration stays the same.

Start calling

Call any model in the key’s scope. Usage, cost and request metadata appear in the console as each request completes.

Read the docs

python · openai sdk

from openai import OpenAI
import os

client = OpenAI(
    base_url="https://models.seedone.ai/v1",
    api_key=os.environ["SEEDONE_API_KEY"],
)

response = client.chat.completions.create(
    model="<model>",
    messages=[{"role": "user", "content": "Hello"}],
)

Pricing

Pay per token, or choose a plan.

Pay from your wallet, or choose a monthly plan that turns a variable bill into a fixed one. Both paths use the same per-token list prices and the same model catalog.

View pricing

Mini

$15/ month

$50 included usage

$3 / 5h
$15 / 7d
$50 / 30d
2 concurrent requests

Get API key

Light

$69/ month

$300 included usage

$15 / 5h
$90 / 7d
$300 / 30d
4 concurrent requests

Get API key

Standard

$119/ month

$600 included usage

$30 / 5h
$180 / 7d
$600 / 30d
8 concurrent requests

Get API key

Premium

$199/ month

$1,200 included usage

$60 / 5h
$360 / 7d
$1,200 / 30d
16 concurrent requests

Get API key

Prices are in USD. Usage is metered per token at the model’s list price. All plans cover the public standard routes.

Security

What we keep, and what you control.

We retain the metadata needed to bill and debug a request. We do not retain its prompt or response body.

Read the security approach

Keys Only you can see it

Full API key values are not available to platform staff, exports or logs.

Payloads Not stored

Prompts and completions are not written to the platform database.

Control Revoked on the next request

Resetting a key invalidates the old value immediately; the action is written to the audit log.

Teams Access you can remove

Member keys stop on removal, and creators never see a member’s full key.

FAQ

Questions before the first request.

Still unsure? Write to support@models.seedone.ai and include a request id if you have one.

Do I need separate keys for the OpenAI and Anthropic protocols?

No. One key works at both entry points. The platform does not translate between protocols; use the protocol that matches the model’s route.

How do I know which models my key can call?

The console lists the models available to that key. The scope is derived from the routes bound to the key, so it can differ between keys and over time.

How is usage metered and billed?

Each request is metered in tokens at the model’s list price. A monthly plan draws down its included quota; pay-as-you-go draws down your wallet. The console shows the final charge for every request.

Can I keep using the official SDKs?

Yes. Use the OpenAI SDK with the OpenAI-compatible Base URL, or the Anthropic SDK with the Anthropic-compatible Base URL. Only the Base URL and key change.

Is registration open?

Registration is invite-only during launch. An invite code is required when you create your account.

What happens if a provider fails?

The gateway retries within a bounded budget and selects another healthy upstream for the same model when one is available. If every attempt fails, you receive an explicit upstream error with a request id.

Why not use each provider directly?

You can. SeedOne is for teams that would rather keep one integration, one key and one bill while the model catalog changes. You trade a platform layer for fewer accounts, SDKs and invoices.