Skip to main content

Two Types of API Keys

In AskTable, there are two types of API keys: API_KEY and LLM_API_KEY. They have different purposes and use cases. Please choose and configure them according to your actual needs.

API Key Description Image

1. API_KEY​

Purpose

Used to call the AskTable backend API, mainly for third-party platforms such as Feishu, Kouzi, or programs you write yourself.

Features

  • Starts with ADMIN_ or ASKER_, for example:
    • ADMIN_JDML38AGS4WKPZLSYPEE
    • ASKER_JDML38AGS4WKPZLSYPEE
  • Different prefixes represent different permissions. For detailed explanations, please refer to: Authentication and Authorization.
Example of API_KEY

Steps to Obtain

  1. Enter project settings: In the AskTable main interface, find the project (e.g., "Public Project") and click the settings button (⚙️) on the right side.
Enter API-Key setting page
  1. Create a new API-Key: Go to the "API-Key" management in the left menu and click "+ New API-Key".
Create a new API-Key

2. LLM_API_KEY​

Purpose

Used to call cloud-based large models (such as DeepSeek, etc.). In the private deployment solution, the system needs this key to access the cloud model to ensure full functionality.

Features

  • Starts with asktable-, for example:
    • asktable-9gwtWZUo1f319A2tGh4GQ1=.....
  • It is written into the docker-compose.yml configuration file via environment variables, with the key name LLM_API_KEY.

Configuration Steps

  1. Enter private deployment configuration: Click the "Private Deployment" button at the top of the AskTable main interface.
Enter private deployment configuration
  1. Configure LLM_API_KEY: In the pop-up window, find "Compute Power → Configure AI Model Token" and apply for or find the LLM_API_KEY.
Configure LLM_API_KEY

Summary: How to Distinguish These Two Types of Keys?​

Key TypePurposePrefix ExampleCommon Scenarios
API_KEYCall AskTable APIADMIN_ / ASKER_Integration with Feishu, Kouzi, third-party applications
LLM_API_KEYCall large models (e.g., DeepSeek)asktable-Private deployment solutions, accessing cloud or local compute power

Note: If connecting to a local large model, set LLM_API_KEY and LLM_BASE_URL to the parameters of your private large model service. Refer to Using a Local LLM Model.