Generic Tool Integration with DataEyesAI (OpenAI-Compatible Protocol)

DataEyesAI's API is fully compatible with the standard OpenAI API protocol. This means that as long as the tool you are using supports a custom OpenAI endpoint (such as Cursor, Chatbox, various VS Code plugins, etc.), you can easily swap the underlying model for DataEyesAI's high-performance compute.

1. The Three Essentials of Universal Configuration

No matter which tool you are using, find its "Settings / API Configuration / Custom Provider" page and replace the following three core parameters:

Parameter	Value	Troubleshooting Tips
API Base URL (or: Endpoint / Server URL)	`https://platform.dataeyes.ai/v1`	Important note about `/v1`: Most tools (such as Cursor) require the `/v1` suffix. However, if you get a "404" or "Invalid URL" error, try removing `/v1` and use only the root domain.
API Key (or: Secret Key)	`sk-xxxxxxxxxxxxxxxx`	Enter the raw plaintext key obtained from the DataEyesAI console. Do not manually add a `Bearer` prefix.
Model (or: Model Name)	The exact name of a model supported by DataEyesAI	The name must match the model listed in the DataEyesAI console exactly, character for character (e.g. `gpt-4o`, `claude-3-5-sonnet-20241022`).

2. Demo: Cursor Integration

Here is a walkthrough using Cursor, one of the most popular AI coding tools:

Note: Due to Cursor's product limitations, the free tier only supports Auto mode and does not allow custom model configuration. To use custom models, please upgrade to Cursor Pro or higher — after which all DataEyesAI-supported models will be available.

Open Cursor and navigate to Settings → Models.

Find the OpenAI configuration section, enable the Override OpenAI Base URL toggle, and enter https://platform.dataeyes.ai/v1.

In the API Key field, enter your DataEyesAI sk-... key.

Click Add Custom Model, enter the name of the model you want to use (e.g. deepseek-v3.2-251201), and enable its toggle.

Select that model in the conversation panel to start using it.

(Example screenshot)

3. Universal Troubleshooting Guide (FAQ)

If you cannot have a conversation after configuring, work through the following three classic errors in order:

1. Error 404 (Not Found) / Invalid URL / Model does not exist?

This is the most common path or name concatenation mistake:

Base URL trailing slash issue: This is caused by differences in how individual tools handle URL construction internally. If adding /v1 causes an error, try removing /v1. If omitting it causes an error, try adding /v1.

Model name issue: Do not guess — make sure the model name you entered actually exists on the DataEyesAI platform and that your account has permission to call it.

2. Error 401 (Unauthorized) / Authentication Failed?

Check whether the key was copied with a trailing space, or whether a Bearer prefix was accidentally added.

3. Error 502/504 / Network Connection Timeout?

This may be caused by temporary routing instability on the primary network. Try replacing the Base URL with an official backup node (e.g. https://platform.dataeyes.ai/v1).

Other Tools

Generic Tool Integration with DataEyesAI (OpenAI-Compatible Protocol)#

1. The Three Essentials of Universal Configuration#

2. Demo: Cursor Integration#

3. Universal Troubleshooting Guide (FAQ)#

Generic Tool Integration with DataEyesAI (OpenAI-Compatible Protocol)

1. The Three Essentials of Universal Configuration

2. Demo: Cursor Integration

3. Universal Troubleshooting Guide (FAQ)