Obtaining API Keys & Authentication

To ensure data security and billing transparency, DataEyesAI physically isolates the API Keys for its two core product lines — "AI Models" and "Search & Reader".

Before making API calls, please familiarize yourself with how to obtain and manage your dedicated keys.

1. AI Model API Key (For LLM, Image, and Video Models)

This key is used to access all generative AI models available in the "Model Marketplace".

1.1 Obtaining and Creating Tokens

The system provides a default "Key" — click the eye icon on the right to view or copy it for immediate use.

For better security control, we recommend clicking Create New Token at the top of the page to generate independent keys for different projects.

1.2 Core Configuration: Group Type & Token Group

When creating a new token, you must select a Group Type. The system supports three intelligent routing strategies:

Price Priority: Dynamically routes to the most cost-effective underlying channel.

Stability Priority: Dynamically routes to the highest-availability (high SLA) channel.

Specified Group: Manually configure the underlying channel with fine-grained control (workflow: select "Model Provider" -> select "Token Group").

Under "Specified Group", group names typically consist of a prefix (model brand or channel source) + suffix (SLA attribute):

Common Prefix Examples (Representing Model Brands or Cloud Channels):

Text & Core LLMs

gpt / openai: OpenAI series

claude / cl / aws: Anthropic Claude series and underlying cloud vendor channels

gemini / gm: Google Gemini series

deepseek / ds: DeepSeek

qwen / qw: Qwen (Alibaba Tongyi Qianwen) series

aliyun / al: Alibaba Bailian series

doubao / db: Doubao (ByteDance) series

glm: Zhipu GLM

hy: Tencent Hunyuan

gr: xAI Grok

Multimodal & Proprietary Models

mj: Midjourney

so / veo: Sora / Google Veo video models

vi / kl / ha / mi / fl: Vidu / Kling / Hailuo / Mimo / Flux

de: Proprietary and self-deployed private models

Core Suffix Descriptions (Representing Underlying SLA & Concurrency Attributes):

-of (Official Channel): 100% native official or tier-1 cloud vendor direct connection. Highest SLA — recommended for enterprise production environments.

-vip (VIP Pool / High-Availability Pool): Premium high-concurrency pool or dedicated exclusive nodes. Ensures peak stability under high concurrency.

-sp (Special Pool / Value Pool): Standard value-tier resource pool. Excellent cost-effectiveness — recommended for development, testing, and daily coding.

-s (Saver / Spot Pool): Maximum value tier (stability not guaranteed). Suitable for cost-sensitive workloads.

-sl / -fa (Slow & Fast): Slow / Fast channels. Dedicated speed tiers for Midjourney and other image generation tasks.

-pro / -cli / -re: Specialized hybrid resource pools for specific account tiers or use cases.

Important Note:
The system dynamically provisions available resources based on your account's permission level and business context. Different accounts may see slight variations in group name length and specific prefixes (e.g., Claude may appear as claude-aws-of or abbreviated as cl-of).
Always refer to the actual group names displayed in your console's dropdown menu. For enterprise-grade high-concurrency or private deployment requirements, please contact your dedicated account manager.

2. Key Security & Advanced Configuration

When creating a new token, DataEyesAI provides a suite of enterprise-grade risk control options. We strongly recommend configuring these before deploying to production:

Quota Limit:
By default, tokens have no quota limit (constrained only by the account's total balance). You can set a fixed USD spending cap for tokens issued to team members or specific tools — once exceeded, the token is automatically deactivated, effectively preventing runaway costs from code loops.

Expiration:
One-click settings for "one month", "one day", or custom expiration times. Ideal for issuing temporary keys for short-term contractor projects or testing.

IP Whitelist:
Bind specific server IP addresses. Even if a key is accidentally leaked, requests from non-whitelisted IPs will be rejected — fundamentally eliminating unauthorized usage.

3. Search & Reader API Key (For Web Search, Web & Document Parsing)

This key provides authentication exclusively for search, page extraction, and document OCR endpoints.

3.1 How to Obtain

Click API Key below it.

Keys generated on this page can only be used to call Search & Reader-specific endpoints such as /v1/search and /v1/reader — they cannot be used for model API calls.

Next Steps

After successfully obtaining your API Key, you can:

Review the Base URL specifications.

Configure your Key and Base URL in your AI coding assistant — see the Developer Tool Integration Guide.

Start making requests programmatically — see the Chat Completions API Reference.

API Key

Obtaining API Keys & Authentication#

1. AI Model API Key (For LLM, Image, and Video Models)#

1.1 Obtaining and Creating Tokens#

1.2 Core Configuration: Group Type & Token Group#

2. Key Security & Advanced Configuration#

3. Search & Reader API Key (For Web Search, Web & Document Parsing)#

3.1 How to Obtain#

Next Steps#