TokenHub · Multi-model AI gateway

One account,
every model you need

DeepSeek, Qwen, GLM and Kimi through one OpenAI-compatible API. One account, one key, one bill — switch models anytime.

Start chatting Download the app

Multi-modelone API

Everywhereweb & desktop

Unifiedbalance & billing

Model router · auto 1.00× ▾ Multi-model · usage-based

Optimize this SQL for deep pagination.

Use cursor pagination so large OFFSETs stay fast:

SELECT *
FROM requests
WHERE id < :last_id
ORDER BY id DESC
LIMIT 20;

Add a composite index on (user_id, id DESC).

Message TokenHub…

Windows Desktop v2 availableDeveloper preview · Android / iOS coming soon

Integrated models

DeepSeekQwenGLMKimiMore coming

DOWNLOAD

Choose your platform

Windows Desktop v2 developer preview is available. Android and iOS are coming soon — one account across all devices.

Android

Android 8.0+ · ~28 MB

Use the web app →

iOS

iOS 15+ · TestFlight

Use the web app →

Windows 10 / 11

Desktop v2.0.0 preview · local workspace · unzip and run

Download Desktop v2 preview SHA256 checksum Use the web app →

Desktop v2 is a standalone local workspace with files, tasks, local tools and permission levels. The preview build is unsigned; the public release will be signed and distributed via MSIX / Microsoft Store.

CAPABILITIES

One account, every device

Switch models freelyAccess leading models in one place and switch mid-conversation without changing accounts.

Smart routing & streamingAutomatic routing by model, streaming responses and usage-based billing.

Unified balance & billingBalance, request history and wallet ledger in one place.

Security & isolationEncrypted upstream keys, isolated accounts and a private admin console.

Start now — no download needed

The web app is ready. Sign in and start chatting with leading models.

Open web chat →

One account,every model you need