API CHECK: Checks the availability of OpenAI-compatible API models.

api-check is a pure front-end API testing tool, open-sourced by developer october-coder on GitHub, which is mainly used to test the usability of API interfaces, especially for OpenAI proxy APIs such as oneapi and newapi. It is mainly used to test the usability of API interfaces , especially support for OpenAI proxy APIs , such as oneapi and newapi. all operations are completed in the browser , does not rely on the back-end server , to ensure data security , while avoiding network timeout problems . The tool provides detailed data such as response time and model consistency, and also supports cloud and local storage test configurations. api-check is easy to operate, supports Chinese and English interfaces and dark and light color modes, and is suitable for developers to quickly verify API performance. It can be deployed via Vercel, Docker or Cloudflare for high flexibility.

Function List

• Test the availability of OpenAI agent APIs such as oneapi and newapi.
• Displays API response times, model consistency, and system fingerprints with clearly visible results.
• Supports cloud storage to save test configurations to the server for multi-device synchronization.
• Supports local storage, caching configurations to the browser for fast loading.
• Provides dark and light color mode switching, adapting to different use environments.
• Supports Chinese and English interfaces to meet different user needs.
• Integrate fast chat testing to validate model responsiveness.
• Supports batch testing of GPT, Claude and Gemini of the key.
• Provide one-click copy function for easy sharing of test results.
• Supports Vercel, Docker and Cloudflare deployment methods.

Using Help

Installation and Deployment

api-check is purely a front-end tool, you can use the online version directly or deploy it yourself. Below are the detailed steps for the three deployment methods:

Vercel deployment

1. Accessing GitHub Repositories
   show (a ticket) https://github.com/october-coder/api-check, go to the project page.
2. One-Click Deployment
   点击页面上的 “Deploy with Vercel” 按钮，或者直接访问 https://vercel.com/new/clone?repository-url=https://github.com/october-coder/api-checkThe
3. Log in and configure
   Log in to Vercel with your GitHub account and add the backend password to the environment variables page, for example:
   • Key:PASSWORD, value:your_passwordThe
4. Completion of deployment
   点击 “Deploy”，几分钟后会生成一个地址，比如 https://api-check-yourname.vercel.appThe
5. Optional operations
   If you need to bind a custom domain name, refer to https://vercel.com/docs/concepts/projects/domains/add-a-domainto avoid the default domain name being restricted in certain regions.

Docker Deployment

1. Run command
   Enter the following command in the terminal for one-click deployment:

docker run -d -p 13000:13000 -e PASSWORD=your_password -v your_path:/app/data --name api-check ghcr.io/rickcert/api-check:latest

2. Parameter description

• -p 13000:13000: Map port to local 13000.
• -e PASSWORD: Set the access password.
• -v your_path:/app/data: Specifies the local storage path.

3. access tool
   Once deployment is complete, open your browser and type http://localhost:13000The

Cloudflare Deployment

1. Reference Tutorial
   interviews https://github.com/october-coder/api-check/blob/main/docs/cloudflare.md, follow the steps.
2. Binding Domain Name
   It is recommended to bind a custom domain name to ensure stable access.

How to use the main features

Testing API Usability

1. Open the api-check page (online version) https://check.crond.dev (or self-deployment address).
2. Enter the test parameters:

• API Key: Fill in the key, e.g. sk-xxxxThe
• URL: Enter the API address, for example https://api.example.comThe
• mould: Select the model, e.g. gpt-4o-miniThe
• timeout: Set the request timeout, e.g. 10 Seconds.
• concurrency: Set the number of simultaneous requests, e.g. 2The

3. 点击 “Test” 按钮，等待结果。界面会显示：

• Response time in milliseconds.
• Model consistency (does it match expectations).
• System fingerprinting (to verify API authenticity).

Saving and loading configurations

• cloud storage：点击 “Save to Cloud”，输入账号和密码，配置会上传到服务器。下次使用时，点击 “Load from Cloud” 加载。
• local storage：点击 “Save Locally”，配置保存到浏览器。下次打开页面自动加载。

Quick Chat Test

1. 在界面找到 “Quick Chat” 选项。
2. 输入测试问题，例如 “1+1等于几？”。
3. Click Send to see the model return results and verify responsiveness and accuracy.
4. transferring entity closeChat: true Disable this feature (suitable for proxy sites).

Batch test key

1. 进入 “Experimental Features” 模块。
2. Enter multiple keys, e.g. GPT Refresh Tokens maybe Claude Session Keys.
3. 点击 “Batch Test”，工具会逐一验证并显示结果。

Advanced Authentication Functions

• Official Agent Verification: Send multiple identical requests to analyze consistency and display system fingerprints.
• Temperature verification: Set the temperature parameter to 0.01, testing model randomness and stability.
• Function call validation: Test whether the model supports function calls and returns the correct result.

Example of operation process

Suppose you want to test an OpenAI agent API:

1. show (a ticket) https://check.crond.devThe
2. Input:

• API Key:sk-test123The
• URL:https://api.test.comThe
• Model:gpt-4oThe
• Timeout:10 Seconds, concurrent:2The

3. 点击 “Test”，结果显示响应时间 300 毫秒，模型一致性通过。
4. 点击 “Save to Cloud”，输入账号保存。
5. 下次打开，点击 “Load from Cloud”，配置自动加载。
6. 进入 “Quick Chat”，输入 “今天天气如何？”，查看返回结果。

This process is simple and straightforward and is good for getting started quickly.

application scenario

1. API Performance Validation
   Developers need to check that the API is stable. api-check shows response times and consistency to help pinpoint problems.
2. Multi-device configuration synchronization
   Teams test APIs on different devices. save configurations with cloud storage that members can load at any time, increasing efficiency.
3. Learning Model Behavior
   Newbies want to understand the patterns of the results returned by the API. Observe model performance through quick chats and temperature validation.

QA

1. Does api-check require backend support?
   Not required. It runs entirely on the front end and data is not uploaded to third party servers.
2. What models are supported?
   Models that primarily support the OpenAI agent API, such as gpt-4o-miniThe API is also compatible with other APIs in similar formats.
3. How can I view the test report?
   Once the test is complete, the interface generates a report with information such as response time, consistency, and fingerprinting.