One Key to call all models DeepSeek · Qwen · GLM · Kimi
One Key for every mainstream model. All models are billed at each vendor's publishedofficial list price. Chat Completions and Messages are both supported — just set a base_url.
Sign up and get ¥10 in free test creditShinLinkAI is operated by Shanghai Xinrui Industrial. It is not affiliated with DeepSeek, Alibaba Cloud, Zhipu AI, Moonshot AI or any other model vendor. Model names are used only to describe compatibility.
API
One gateway for mainstream AI coding tools and production APIs.
Choose a top-up packOverview
What is ShinLinkAI?
ShinLinkAI aggregates models from several upstream vendors behind one Chat Completions / Messages compatible API. The same balance works for coding agents, IDEs, SDKs and production — each Key can be rate-limited separately, and usage is tracked precisely down to the token.
API protocols
Messages /v1/messages, Chat Completions /v1/chat/completions, and coding tools use Responses.
Built for tools
Point a custom base_url at any tool — Cursor, Cherry Studio, major IDEs, Cherry Studio, VS Code extensions and any OpenAI-compatible SDK.
One balance
Top-up credit never expires. Each request is billed precisely by input, output, cached and reasoning tokens.
Set up in minutes
Connect once, use everywhere
Any tool that supports a custom base_url just needs the ShinLinkAI address. No separate account per model, and no overseas credit card.
Paste the ready-made config block into your tool and it works.
Create a Key
Generate an sk-tp Key in the console.
Pick a tool
Cursor, Cherry Studio, a major IDE, or any OpenAI-compatible SDK.
Enter the URL
https://api.shinloo.com/v1 — start calling.
One Key, every vendor
DeepSeek for reasoning, DeepSeek for code, Qwen/Kimi/GLM for localisation — routed automatically by model name.
Multi-protocol
Messages, Chat Completions and Responses all draw on the same balance; change the base_url with zero code changes.
Streaming ready
SSE streaming output, suited to coding agents / IDEs / production calls, at the same price as ordinary requests.
Controllable spend
Each Key has its own limit, and request history plus token-level usage detail is always available.
Native experience
Model calls, as smooth as the official ones.
The same gateway Key serves agents, IDE extensions and production APIs. Simple setup, transparent billing, every Key under control.
One Key, every frontier model
DeepSeek · Qwen · GLM · Kimi — flagship and mainstream models on the same balance.
K3 flagship with strong agentic and long-document performance.
GLM-5.3 top-tier coding gains; 5.3 Flash cheap and quick.
DeepSeek-V4-Pro competes with top closed models; DeepSeek-V4-Flash is efficient and affordable. Native 1M context.
Qwen3.8-Max, a 2.4-trillion open flagship; Flash activates 6B with a 1M context; VL for vision.
Simple, transparent usage-based billing
Top up once and you are only charged when you make a request; from ¥100, or any custom amount. Every call has token-level detail in the console.
Personal development and small projects; try every model.
Daily agent use and coding calls.Popular
Active teams and production workloads.
Heavy usage and a team standard; credit never expires.
Enterprise customers and high-concurrency production; VAT invoice available.
Sign-up bonus ¥10 in test credit — verify the gateway before you top up.
How billing works
Streaming and non-streaming cost the same.
Cache hits cost less: as low as ¥0.30 per million (about 1/30 of the input price) — save about 97% on repeated prefixes
FAQ
The questions people usually ask before buying a Key.