Using ChatGPT, Claude, and Gemini

You can drive Aigeon from any AI assistant that supports function calling or tool use. The pattern is the same regardless of the model: define Aigeon endpoints as tools, let the model decide when to call them, and execute the calls against the REST API.

OpenAI / ChatGPT (function calling)

Define the tools in your assistant configuration:

{
  "tools": [
    {
      "type": "function",
      "function": {
        "name": "aigeon_add_contacts",
        "description": "Add one or more contacts to an Aigeon list",
        "parameters": {
          "type": "object",
          "properties": {
            "group_id": { "type": "string" },
            "members": {
              "type": "array",
              "items": {
                "type": "object",
                "properties": {
                  "email": { "type": "string" },
                  "first_name": { "type": "string" },
                  "tags": { "type": "array", "items": { "type": "string" } }
                },
                "required": ["email"]
              }
            }
          },
          "required": ["group_id", "members"]
        }
      }
    }
  ]
}

When the model returns a tool_call, execute it:

async function executeTool(name: string, args: Record<string, unknown>) {
  if (name === "aigeon_add_contacts") {
    const res = await fetch(
      "https://api.aigeon.ai/public/api/v1/contacts/add-group-members",
      {
        method: "POST",
        headers: {
          "X-Api-Token": process.env.AIGEON_API_KEY!,
          "Content-Type": "application/json",
        },
        body: JSON.stringify(args),
      },
    );
    return res.json();
  }
}

Anthropic / Claude (tool use)

Define the same tool in Claude's format and handle tool_use content blocks:

import Anthropic from "@anthropic-ai/sdk";

const client = new Anthropic();

const response = await client.messages.create({
  model: "claude-opus-4-7",
  max_tokens: 1024,
  tools: [
    {
      name: "aigeon_send_flow",
      description: "Trigger an Aigeon flow for a contact",
      input_schema: {
        type: "object" as const,
        properties: {
          flow_id: { type: "string" },
          contact_email: { type: "string" },
          context: { type: "object" },
        },
        required: ["flow_id", "contact_email"],
      },
    },
  ],
  messages: [
    {
      role: "user",
      content: "Trigger the welcome flow for alice@example.com with order_id ord_999",
    },
  ],
});

for (const block of response.content) {
  if (block.type === "tool_use" && block.name === "aigeon_send_flow") {
    await fetch(
      "https://api.aigeon.ai/public/api/v1/flow/send-email-through-flow",
      {
        method: "POST",
        headers: {
          "X-Api-Token": process.env.AIGEON_API_KEY!,
          "Content-Type": "application/json",
        },
        body: JSON.stringify(block.input),
      },
    );
  }
}

Google Gemini (function calling)

import google.generativeai as genai
import requests, os

genai.configure(api_key=os.environ["GOOGLE_API_KEY"])

add_contacts_fn = genai.protos.FunctionDeclaration(
    name="aigeon_add_contacts",
    description="Add contacts to an Aigeon list",
    parameters=genai.protos.Schema(
        type=genai.protos.Type.OBJECT,
        properties={
            "group_id": genai.protos.Schema(type=genai.protos.Type.STRING),
            "emails": genai.protos.Schema(
                type=genai.protos.Type.ARRAY,
                items=genai.protos.Schema(type=genai.protos.Type.STRING),
            ),
        },
        required=["group_id", "emails"],
    ),
)

model = genai.GenerativeModel(
    "gemini-2.5-pro",
    tools=[add_contacts_fn],
)
chat = model.start_chat()
response = chat.send_message("Add alice@example.com to list sg_beta_testers")

for part in response.parts:
    if fn := part.function_call:
        if fn.name == "aigeon_add_contacts":
            members = [{"email": e} for e in fn.args["emails"]]
            requests.post(
                "https://api.aigeon.ai/public/api/v1/contacts/add-group-members",
                headers={"X-Api-Token": os.environ["AIGEON_API_KEY"]},
                json={"group_id": fn.args["group_id"], "members": members},
            )

Simplest path: MCP

If you use Claude Desktop, Claude Code CLI, or Cursor, the MCP server handles all of the above automatically — no tool definitions to write, no execution logic. Just describe what you want and the MCP server maps it to the correct Aigeon API call.

For code-writing agents: Agent Skills

If your AI is writing code that calls the Aigeon API (as opposed to calling it live from a chat), install the aigeon-skills plugin instead of (or alongside) any of the above. It ships the payload shapes, auth rules, scope guidance, and common-pitfall notes the agent needs to produce correct code on the first try. See Agent Skills.