Skip to main content

​
Codex++ Agnes Model Configuration Guide

​
1. Overview

This guide explains how to configure the Agnes model in Codex++. After configuration, you can use the Agnes model directly in Codex++ for chat, code generation, and agent tasks without logging in to a GPT account.

​
2. Prerequisites

Before you start, make sure you have:
  1. An Agnes API Key.
  2. A network connection that can access the Agnes API Gateway.
  3. The target model name.
  4. Codex++ downloaded or ready to install.
Recommended model: agnes-2.0-flash API Base URL: https://apihub.agnes-ai.com/v1

​
3. Download Codex++

Open the Codex++ download page: https://github.com/BigPizzaV3/CodexPlusPlus You can also open the latest release page directly: https://github.com/BigPizzaV3/CodexPlusPlus/releases/latest Download the version that matches your device: If you are not sure whether your Mac uses an Intel chip or an Apple chip, click the Apple icon in the top-left corner and select About This Mac.

​
4. Install Codex++

After downloading, install Codex++ normally. If you are using macOS and see the following error:
Open Terminal and run the following commands:
After running the commands, reopen the Codex++ Management Tool.

​
5. Open the Codex++ Management Tool

After installation, open:
In the left sidebar, click:
Then click:
Next, fill in the Agnes provider information according to the configuration below.

​
6. Add the Agnes Provider

Fill in the following fields: Configuration example:

​
7. Configuration Notes

Please check the following carefully:
  1. Only enter the API key starting with sk-. Do not include Bearer.
  2. The Base URL should only end with /v1. Do not enter /chat/completions.
  3. The upstream protocol must be Chat Completions.
  4. The model name should be:
agnes-2.0-flash
  1. After completing the configuration, click:

​
8. Enable the Agnes Provider

After saving the configuration, return to the provider list and select:
If you see any of the following buttons, click it to enable the Agnes provider:
After it is enabled successfully, Agnes will become the current active provider.

​
9. Start Codex++

Return to the left sidebar and click:
Then click:
You can also click the button in the top-right corner:
After startup, you can use Codex++ normally, and requests will be routed through the Agnes model.

​
10. Verify the Configuration

Open Codex++ and create a new session. Enter a simple test message, for example:
If Codex++ responds normally, the Agnes model configuration is successful.

​
11. FAQ

​
1. What should I do if Codex does not respond properly?

Please check the following:
  • Make sure you started Codex from Codex++ or the Codex++ Management Tool, instead of opening the original Codex.
  • Make sure the Base URL is:
https://apihub.agnes-ai.com/v1
  • Make sure the Key only contains the API key starting with sk-.
  • Make sure the upstream protocol is:
  • Make sure the model is:
agnes-2.0-flash
  • Make sure the Agnes provider has been enabled after saving the configuration.

​
2. How should I enter the API Key?

Only enter the Agnes API Key itself, for example:
Do not enter:

​
3. How should I enter the Base URL?

Correct: https://apihub.agnes-ai.com/v1 Incorrect: https://apihub.agnes-ai.com/v1/chat/completions

​
4. Do I need to log in to a GPT account?

No. After configuration, you can use Codex++ directly with the Agnes model. Provider Name: Agnes Mode: API Only Model: agnes-2.0-flash Base URL: https://apihub.agnes-ai.com/v1 API Key: YOUR_AGNES_API_KEY Protocol: Chat Completions

​
13. Full Setup Summary

  1. Download and install Codex++.
  2. Open the Codex++ Management Tool.
  3. Go to Provider Configuration.
  4. Add the Agnes provider.
  5. Enter the model, Base URL, API Key, and upstream protocol.
  6. Save the configuration.
  7. Enable the Agnes provider.
  8. Start or restart Codex++.
  9. Create a new session to test whether Codex++ responds normally.
  10. After configuration, you can use the Agnes model in Codex++.