THE CHINESE ONE
DEPTH 01 · THE THRONE ROOM
Dithered 1-bit scene: the old Mage on his chair points a lazy finger; the hunched Servant bows, eager to obey.
ERR 429 Claude refused your request, m'lord. Weekly limit reached · resets Thu 09:00
MESSAGE LOG

SERVANT ›your agent Yes m'lord.Yes m'lord.

One key to US-hosted DeepSeek, Qwen, Kimi, GLM and MiniMax, inside the agent you already use. Your agent hands the grunt work to cheap open models and keeps your Claude plan for the hard parts.

see the plans

US-hosted · code never touches China · from $9/mo · refusal-free models, if you need them

Three rooms. Then the grunt work never touches your plan.

ROOM 01 · The Key Vault

One key.

DeepSeek, Qwen, Kimi, GLM and MiniMax behind a single key. Hosted in the US, so your code never touches China.

ROOM 02 · The Command Chamber

One command.

npx thechineseone init

Hands the key to your servant, the agent you already use: Claude Code, Cursor, Cline, Hermes or OpenClaw. First call in under two minutes.

ROOM 03 · The Workshop

Your Servant, the agent, typing fast on a tiny keyboard as it summons the Chinese one; a pile of finished scrolls beside it.

The Chinese one does the grunt work.

Your servant summons it for tests, renames, docs, boilerplate and migrations. Your Claude or GPT plan stays fresh for the hard parts.

up to ~40% of your weekly session back*

*An estimate for grunt-heavy weeks, not a promise. Your savings depend on how much work you delegate.

ROOM 04 · The Summoning Circle

Pick who to summon. The cheap one is the default.

Eight open models, one key, all hosted in the US. Each sheet shows what a heavy agent session costs on it. Your servant picks the cheap one unless you say otherwise.

Launch prices in USD, may change. Estimate for a heavy agent session: 2 to 3 hours of agentic coding, about 15M input tokens (90% cached) and 0.2M output. Not a promise. Plans and the weekly limit are on the pricing page.