AI Agents & MCPs & AI Workflow Automation • (~400 MCP servers for AI agents) • AI Automation / AI Agent with MCPs • AI Workflows & AI Agents • MCPs for AI Agents
TypeScript ai-agentai-agent-toolsai-agentsai-agents-frameworkmcpmcp-servermcp-toolsmcpsn8n-alternativeno-code-automationworkflowworkflow-automationworkflows
import { createAction, Property } from '@activepieces/pieces-framework';
import {
AuthenticationType,
HttpMethod,
HttpRequest,
httpClient,
propsValidation,
} from '@activepieces/pieces-common';
import { cometApiAuth } from '../common/auth';
import * as z from 'zod/mini'
import { modelIdDropdown } from '../common/props';
interface CometApiChatResponse {
id: string;
object: string;
created: number;
model: string;
choices: {
index: number;
message: {
role: string;
content: string;
};
finish_reason: string;
}[];
usage: {
prompt_tokens: number;
completion_tokens: number;
total_tokens: number;
};
}
interface ChatMessage {
role: 'system' | 'user' | 'assistant';
content: string;
}
export const askCometApiAction = createAction({
audience: 'both',
name: 'ask-cometapi',
classification: 'READ',
displayName: 'Ask CometAPI',
description: 'Sends a prompt to any AI model supported by CometAPI.',
aiMetadata: {
description:
'Sends a single-turn chat completion (one prompt plus an optional system message, with no multi-turn history) to any model in the CometAPI catalog, which fronts OpenAI, Anthropic, Gemini and other vendors behind one key. Requires a model id chosen from CometAPI\'s live model list, and accepts temperature, top_p and max tokens as optional sampling controls; pick it when the model should be selectable at runtime rather than fixed to a single vendor, and use this piece\'s Custom API Call for any other CometAPI endpoint such as image generation or embeddings. Not idempotent: each call runs a fresh generation and consumes tokens.',
idempotent: false,
},
auth: cometApiAuth,
props: {
model: modelIdDropdown,
prompt: Property.LongText({
displayName: 'Prompt',
required: true,
description: 'The prompt to send to the AI model.',
}),
systemMessage: Property.LongText({
displayName: 'System Message',
required: false,
description: 'Optional system message to set the behavior of the AI.',
}),
temperature: Property.Number({
displayName: 'Temperature',
required: false,
description:
'Controls randomness (0-2). Lower values make output more focused.',
defaultValue: 1.0,
}),
maxTokens: Property.Number({
displayName: 'Max Tokens',
required: false,
description: 'Maximum number of tokens to generate.',
defaultValue: 1000,
}),
topP: Property.Number({
displayName: 'Top P',
required: false,
description: 'Controls diversity via nucleus sampling (0-1).',
defaultValue: 1.0,
}),
},
async run(context) {
// Validate input parameters
await propsValidation.validateZod(context.propsValue, {
temperature: z.optional(z.number().check(z.minimum(0), z.maximum(2))),
topP: z.optional(z.number().check(z.minimum(0), z.maximum(1))),
maxTokens: z.optional(z.number().check(z.positive())),
});
const { model, prompt, systemMessage, temperature, maxTokens, topP } =
context.propsValue;
// Build messages array
const messages: ChatMessage[] = [];
if (systemMessage) {
messages.push({
role: 'system',
content: systemMessage,
});
}
messages.push({
role: 'user',
content: prompt,
});
// Build request body
const requestBody: Record<string, unknown> = {
model: model,
messages: messages,
};
// Add optional parameters
if (temperature !== undefined) requestBody['temperature'] = temperature;
if (maxTokens !== undefined) requestBody['max_tokens'] = maxTokens;
if (topP !== undefined) requestBody['top_p'] = topP;
// Build request
const request: HttpRequest = {
url: 'https://api.cometapi.com/v1/chat/completions',
method: HttpMethod.POST,
body: requestBody,
authentication: {
type: AuthenticationType.BEARER_TOKEN,
token: context.auth.secret_text,
},
headers: {
'Content-Type': 'application/json',
},
};
try {
const response = await httpClient.sendRequest<CometApiChatResponse>(
request
);
const responseContent = response.body.choices[0]?.message?.content;
if (!responseContent) {
throw new Error('No response content received from API');
}
return {
response: responseContent.trim(),
model: response.body.model,
usage: response.body.usage,
};
} catch (error: unknown) {
// Detailed error handling
const errorObj = error as {
response?: { status?: number; body?: { error?: { message?: string } } };
message?: string;
};
if (errorObj.response?.status === 401) {
throw new Error('Authentication failed: Please check your API key.');
}
if (errorObj.response?.status === 429) {
throw new Error('Rate limit exceeded: Please try again later.');
}
if (errorObj.response?.status === 400) {
const errorMessage =
errorObj.response?.body?.error?.message || 'Bad request parameters';
throw new Error(`Request parameter error: ${errorMessage}`);
}
if (errorObj.response?.status === 404) {
throw new Error(`Model not found: ${model}`);
}
if (errorObj.response?.status && errorObj.response.status >= 500) {
throw new Error('Server error: Please try again later.');
}
throw new Error(`Request failed: ${errorObj.message || 'Unknown error'}`);
}
},
});