This document explains how to use Newcclm. If you are using the command line client, seeClient Guide; If you want to integrate Newcclm into your own tools, see the "Client Integration" section below.
This guide explains how to integrate Newcclm into your tool, or directly use the command line client we provide.
After purchasing you will get aAkun(username + password) -
The command line client uses it to log in; if you want to access your own tools, another one will be sentsk-*format.
The protocol always usesOpenAI compatible. Don’t choose Anthropic native—— We are using the OpenAI set of interfaces. If you choose the wrong one, you will not be able to connect.
Before you get started, make sure you have:
https://sccclm.com/v1's network environment.No matter which tool you pick up, you only need these few items. Can be copied directly.
sk-*Key
gpt-4o、claude-*The name of this type of model. We only recognize our own stall name (see the table above), and if you fill in anything else, it will be reported that the model does not exist./v1。If the tool will fill in the path by itself, you only need to fill in thehttps://sccclm.comThat’s it – see the separate pages below for the differences between the tools.Before configuring the client, you can use the following to confirm whether the key and model are available.
PutYOUR_API_KEYReplace it with your real key, paste it in the terminal and execute:
curl https://sccclm.com/v1/chat/completions \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "newcclm-fast",
"messages": [{"role": "user", "content": "Hello"}]
}'Just return a piece of JSON with content. If you reportInvalid tokenwhen writing code to check whether the key has been copied completely;
Newspapermodel not found, check the spelling of the model name.
Command line clientKey not required—— Log in directly with your account. The key is only needed if you want to connect Newcclm to your own tools such as Cursor and WorkBuddy. The version of
If you need a key, contact us and explain which tool you want to connect, and we will send it to you. The key format issk-The first string of characters.
Select the tool you are using and configure it according to the corresponding guide.
our own command line client. A command is installed and can be used once logged in. Key not required。
Configuration Tutorials →A desktop chat client with a Chinese interface. Just create a new custom provider.
Configuration Tutorials →There is a desktop version as well as a mobile version. You can see this mainly on mobile phones.
Configuration Tutorials →Use this if you don't want to install any software. Fill in three values in the browser to chat.
Configuration Tutorials →Just overwrite the OpenAI interface address. It is recommended to usenewcclm-coding。
Add Custom Provider, Agreement Selectionopenai-completions. Don’t think of it as DeepSeek’s official access.
Open source terminal encoding agent. Add a custom provider to point to our address.
Configuration Tutorials →Click a link and the model will appear inClaude Code / Codex / Pi / OpenCode / Cursorinside. Suitable for people who have installed multiple Agents.
Access instructions →Any tool that supports custom OpenAI compliant interfaces can be connected. Not on it? Look at this generic fill.
General guidelines →For OpenAI compatible services, the Base URL usually ends with/v1End
# Generally fill this
https://sccclm.com/v1
# Unless the client explicitly requires the complete endpoint (this is the case for a few tools)
https://sccclm.com/v1/chat/completions
# Don't do this: it will stack into /v1/v1/...
https://sccclm.com/v1/v1Some tools will ask you "use OpenAI or Anthropic format." Always choose OpenAI. If you choose Anthropic, you will be unable to connect directly.
| Can be filled in | Don't fill in |
|---|---|
newcclm（日常）newcclm-fast&Fastnewcclm-codingTingkatkannewcclm-free（免费） |
gpt-4o、claude-3-5-sonnet、deepseek-chatWait for the real model name |
Related pages: Home · Price · FAQs · Query Usage