.do

Capabilities

Skills, tools, and abilities available to AI agents

Agent Capabilities

Agents on the .do platform have access to a wide range of capabilities spanning natural language, code, data, and integration with external systems.

Core Capabilities

Natural Language Processing

Understanding

  • Intent classification
  • Entity extraction
  • Sentiment analysis
  • Context interpretation
  • Multi-turn conversation

Generation

  • Content creation (articles, emails, documents)
  • Code generation and explanation
  • Translation and localization
  • Summarization and synthesis
  • Creative writing

Code & Development

Code Generation

  • Write functions and modules
  • Generate tests and documentation
  • Refactor and optimize code
  • Fix bugs and security issues

Code Execution

  • Run Python, JavaScript, TypeScript
  • Execute SQL queries
  • Shell commands (sandboxed)
  • API testing and validation

Development Tools

  • Git operations
  • Package management
  • Build and deployment
  • Code review and analysis

Data Processing

Extraction

  • Parse structured data (JSON, CSV, XML)
  • Extract from unstructured text
  • OCR and document processing
  • Web scraping and crawling

Transformation

  • Data cleaning and normalization
  • Format conversion
  • Aggregation and summarization
  • Enrichment from external sources

Analysis

  • Statistical analysis
  • Pattern recognition
  • Anomaly detection
  • Predictive modeling

Integration Capabilities

APIs & Services

Agents can integrate with 1000s of APIs:

Business Systems

  • CRM (Salesforce, HubSpot)
  • ERP (NetSuite, SAP)
  • Accounting (QuickBooks, Xero)
  • HR (BambooHR, Workday)

Communication

  • Email (Gmail, Outlook)
  • Messaging (Slack, Teams)
  • SMS (Twilio, MessageBird)
  • Video (Zoom, Meet)

Developer Tools

  • Version Control (GitHub, GitLab)
  • CI/CD (Jenkins, CircleCI)
  • Monitoring (Datadog, New Relic)
  • Cloud Platforms (AWS, GCP, Azure)

Database Operations

Direct access to data storage:

// Query data
const customers = await db.Customer.query({
  where: { status: 'active' },
  limit: 100,
})

// Update records
await db.Order.update(orderId, {
  status: 'fulfilled',
  fulfilledAt: new Date(),
})

// Create relationships
await db.Customer.relate(customerId, $.hasOrders, orders)

File Operations

Handle documents and files:

  • Read/write local and cloud storage
  • PDF generation and parsing
  • Image processing and generation
  • Spreadsheet manipulation
  • Archive handling (zip, tar)

Advanced Capabilities

Multi-Modal Processing

Vision

  • Image analysis and classification
  • Object detection
  • OCR and document understanding
  • Image generation and editing

Audio

  • Speech-to-text transcription
  • Text-to-speech synthesis
  • Audio analysis and classification
  • Music generation

Reasoning & Planning

Logical Reasoning

  • Deductive and inductive reasoning
  • Constraint solving
  • Optimization problems
  • Game theory and strategy

Planning

  • Task decomposition
  • Resource allocation
  • Scheduling and sequencing
  • Contingency planning

Learning & Adaptation

From Feedback

  • User corrections and preferences
  • Success/failure outcomes
  • Performance metrics
  • A/B test results

From Experience

  • Pattern recognition over time
  • Anomaly detection improvements
  • Personalization
  • Workflow optimization

Capability Composition

Combine capabilities for complex tasks:

const agent = $.Agent.create({
  name: 'Market Research Agent',
  capabilities: [
    // Data gathering
    'web-search',
    'web-scraping',
    'api-integration',

    // Analysis
    'data-extraction',
    'sentiment-analysis',
    'trend-analysis',

    // Output
    'report-generation',
    'visualization',
    'presentation',
  ],
})

Capability Marketplace

Browse and install capabilities from the marketplace:

  • Pre-built tool integrations
  • Industry-specific skills
  • Custom capabilities from community
  • Certified enterprise capabilities

Building Custom Capabilities

Extend agents with custom tools:

const customTool = $.Tool.create({
  name: 'inventory-checker',
  description: 'Check product inventory levels',
  parameters: {
    sku: { type: 'string' },
    warehouse: { type: 'string', optional: true },
  },
  execute: async ({ sku, warehouse }) => {
    // Your custom logic
    return inventoryData
  },
})

agent.addTool(customTool)

Next Steps

  • Teams - Coordinate multiple agents
  • Services - Package capabilities as services
  • Examples - See capabilities in action