Quick Start

Get up and running with the AgenticWork SDK in minutes.

Basic Chat Completion

import { AgenticWork } from '@agentic-work/sdk';

const aw = new AgenticWork({
  apiEndpoint: 'https://api.agenticwork.io',
  authToken: process.env.AGENTICWORK_TOKEN
});

await aw.init();

const response = await aw.complete({
  model: 'gpt-4o',
  messages: [
    { role: 'system', content: 'You are a helpful assistant.' },
    { role: 'user', content: 'What is the capital of France?' }
  ]
});

console.log(response.choices[0].message.content);
// Output: The capital of France is Paris.

Streaming Responses

const stream = aw.streamComplete({
  model: 'claude-3-5-sonnet-20241022',
  messages: [{ role: 'user', content: 'Write a haiku about coding' }]
});

for await (const chunk of stream) {
  if (chunk.type === 'text_delta') {
    process.stdout.write(chunk.text);
  }
}

Using Tools

// Get all available MCP tools
const tools = aw.getMCPTools();
console.log(`Available tools: ${tools.length}`);

// Use tools in completion
const response = await aw.complete({
  model: 'gpt-4o',
  messages: [{ role: 'user', content: 'Search for the latest AI news' }],
  tools: tools.filter(t => t.name.startsWith('awp_web'))
});

Creating an Agent

Agents automatically handle tool execution in a loop:

const agent = aw.createAgent({
  model: 'gpt-4o',
  systemPrompt: 'You are a coding assistant.',
  tools: aw.getMCPTools(),
  maxIterations: 10
});

// Streaming agent execution
for await (const text of agent.run({
  prompt: 'Create a Python function that calculates fibonacci numbers'
})) {
  process.stdout.write(text);
}

Direct Provider Access

Skip the platform and connect directly to providers:

import { createProvider } from '@agentic-work/sdk/providers';

// OpenAI
const openai = createProvider({
  type: 'openai',
  apiKey: process.env.OPENAI_API_KEY
});

// Anthropic
const anthropic = createProvider({
  type: 'anthropic',
  apiKey: process.env.ANTHROPIC_API_KEY
});

// Ollama (local)
const ollama = createProvider({
  type: 'ollama',
  baseUrl: 'http://localhost:11434'
});

// Use any provider with the same API
const response = await openai.complete({
  model: 'gpt-4o',
  messages: [{ role: 'user', content: 'Hello!' }]
});

Working with Flowise Workflows

// List workflows
const workflows = await aw.listWorkflows();

// Create a workflow
const workflow = await aw.createWorkflow({
  name: 'Data Processor',
  description: 'Process and analyze data',
  flowData: { /* Flowise flow definition */ }
});

// Execute workflow
const result = await aw.executeWorkflow(workflow.id, {
  input: 'Process this data'
});

Managing Workspaces

// List workspaces
const workspaces = await aw.listWorkspaces();

// Create workspace
const workspace = await aw.createWorkspace('My Project');
console.log(`Created workspace: ${workspace.id}`);

Checking Permissions

// Check user info
console.log(`User: ${aw.user?.email}`);
console.log(`Admin: ${aw.user?.isAdmin}`);

// Check specific permissions
if (aw.can('manage_workflows')) {
  // User can manage workflows
}

// Check group membership
const platform = aw.getPlatform();
if (platform.isInGroup('developers')) {
  // User is in developers group
}

Error Handling

try {
  const response = await aw.complete({
    model: 'gpt-4o',
    messages: [{ role: 'user', content: 'Hello' }]
  });
} catch (error) {
  if (error.message.includes('rate limit')) {
    console.log('Rate limited, please retry');
  } else if (error.message.includes('authentication')) {
    console.log('Invalid credentials');
  } else {
    console.error('Error:', error.message);
  }
}

Next Steps

• API Reference - Complete API documentation

• Agents - Building agentic applications

• Tools - Working with MCP tools

• Streaming - Advanced streaming patterns