Custom Agents

Cencori's Gateway exposes a standard OpenAI-compatible API at https://cencori.com/api/v1. This means any tool that lets you set a custom OpenAI base URL can connect to Cencori.

Quick Start

1. Deploy from the Marketplace

Navigate to Agents → Agent Marketplace in your project dashboard and click Deploy Agent on the Custom Agent card.

2. Generate an API Key

Go to the agent's Configuration tab and click Generate Key.

3. Configure Your Tool

Look for settings in your tool related to "Custom API Endpoint", "OpenAI Proxy", or "Base URL". Set them as follows:

• API Base: https://cencori.com/api/v1
• API Key: cake_YOUR_KEY_HERE
• Model Name: Use any model ID supported in your Cencori dashboard (e.g., gpt-4o, claude-3-5-sonnet-20240620).

[!NOTE] Some tools append /chat/completions automatically, while others expect you to include it in the base URL. If connections fail, try setting the URL to https://cencori.com/api/v1/chat/completions.

Testing the Connection

You can verify your agent is connected correctly by sending a raw curl request:

curl -X POST https://cencori.com/api/v1/chat/completions \
  -H "Authorization: Bearer cake_YOUR_KEY_HERE" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4o",
    "messages": [
      {
        "role": "system",
        "content": "You are connected."
      },
      {
        "role": "user",
        "content": "Hello!"
      }
    ],
    "stream": false
  }'

This request will appear instantly in your Cencori Logs tab.

Popular Tool Configurations

Cline / Continue.dev (VS Code)

In your IDE extension settings:

1. Select "OpenAI Compatible" as the provider
2. Base URL: https://cencori.com/api/v1
3. API Key: Your cake_ key
4. Model ID: e.g., claude-3-5-sonnet-20240620

Aider

Run Aider from the terminal using standard OpenAI environment variables:

export OPENAI_API_BASE=https://cencori.com/api/v1
export OPENAI_API_KEY=cake_YOUR_KEY_HERE
aider --openai-model gpt-4o

LangChain

from langchain_openai import ChatOpenAI
 
llm = ChatOpenAI(
    openai_api_base="https://cencori.com/api/v1",
    openai_api_key="cake_YOUR_KEY_HERE",
    model_name="gpt-4o"
)