Make

In the new boilerplate, you can create AI Agents with a Low-Code solution using Make API

Environment variables

1. Log in to your Make account

2. Copy your company ORGANIZATION ID and paste in .env

.env
MAKE_ORGANIZATION_ID=

3. Copy your TEAM ID and paste in .env

.env
MAKE_TEAM_ID=

4. Create your API KEY here and paste in .env

.env
MAKE_API_KEY=

5. Copy your MAKE URL, add a /api/v2 and paste in .env

.env
MAKE_API_URL=https://eu2.make.com/api/v2

Scenario templates

1. Create a new scenario in Make

2. Add your scenario data to the scenarios/page.tsx availableScenarios like this:

scenarios/page.tsx
const availableScenarios: ScenarioTemplate[] = [
	{
		id: yourScenarioId,
		image: '/yourScenarioImage.png',
		name: 'yourScenarioName',
		description: 'yourScenarioDescription',
		fields: ['yourScenarioFields'],
		route: '/api/scenarios/yourScenarioName',
	},
]

3. Create a new route in app/api/scenarios. You must specify each connection and the initial webhook, if any. Like this:

• Connection is created like this:

app/api/scenarios/yourScenarioName.ts
// 1. Create OpenAI connection
const connectionResponse = await axios.post(
	'BASE_URL/connections?teamId=MAKE_TEAM_ID',
	{
		accountName: 'OpenAI date',
		accountType: 'openai-gpt-3',
		apiKey: apiKey,
		apiOrg: apiOrg,
	},
	{ headers: { Authorization: 'Token MAKE_API_KEY' } }
)
const connectionId = connectionResponse.data.connection.id
if (!connectionId) {
	throw new Error('Connection ID missed')
}

• Webhook is created like this:

app/api/scenarios/yourScenarioName.ts
// 2. Create webhook for receiving messages
const webhookResponse = await axios.post(
	'BASE_URL/hooks?teamId=MAKE_TEAM_ID',
	{
		name: 'webhook openAI date',
		team_id: MAKE_TEAM_ID,
		teamId: MAKE_TEAM_ID,
		typeName: 'gateway-webhook',
		method: true,
		headers: true,
		stringify: false,
		interface: [{ name: 'message', type: 'text' }],
	},
	{ headers: { Authorization: 'Token MAKE_API_KEY' } }
)
const webhookId = webhookResponse.data.hook.id
const webhookLink = webhookResponse.data.hook.url
if (!webhookId || !webhookLink) {
	throw Error('hook creating error')
}

4. GET a blueprint of your scenario from /scenarios/SCENARIO_ID/blueprint

app/api/scenarios/yourScenarioName.ts
// 3. Get the blueprint of the source scenario
const blueprintResponse = await axios.get(
	'BASE_URL/scenarios/scenarioId/blueprint',
	{
		headers: { Authorization: 'Token MAKE_API_KEY' },
	}
)
const blueprint = blueprintResponse.data.response.blueprint
const flow = blueprint.flow as FlowModule[]

5. Update a blueprint of your scenario to new connections and webhooks

app/api/scenarios/yourScenarioName.ts
// 4. Update module configurations in the flow
const updatedFlow = flow.map(module => {
	// Update webhook module with new webhook ID
	if (module.module === 'gateway:CustomWebHook') {
		return {
			...module,
			parameters: {
				...module.parameters,
				hook: webhookId,
			},
		}
	}
	// Update OpenAI module with new connection and assistant ID
	if (module.module === 'openai-gpt-3:messageAssistantAdvanced') {
		return {
			...module,
			parameters: {
				...module.parameters,
				__IMTCONN__: connectionId,
			},
			mapper: {
				...module.mapper,
				assistantId: assistantId,
			},
		}
	}
	return module
})

6. Create new scenario using new blueprint and synchronize it with Database

Interaction with functionality