functions.do/readme.md

A framework and platform for easily creating, managing, evaluating, and iterating on AI functions & object generation

# functions.do

AI is transforming businesses but integrating LLMs into existing systems presents challenges due to the clash between AI's non-deterministic nature and traditional software's deterministic characteristics.

Key challenges include:
- **Reliability**: AI's unpredictable outputs complicate testing and maintenance
- **Accuracy**: Models can hallucinate or produce incorrect information
- **Model Selection**: Balancing capabilities, speed, and cost across rapidly evolving models
- **Prompt Engineering**: More art than science, difficult to standardize
- **Configuration**: Complex parameter interactions require careful tuning

### The Solution

functions.do creates a clean separation between AI capabilities and application code through strongly-typed interfaces that hide model complexities, enabling:
- Rapid prototyping of AI applications
- Continuous improvement without disrupting application code
- Comprehensive evaluation and optimization strategies


### Usage

You can use it by simply calling any function name with any arguments on the `ai` object like:

```typescript
import { ai } from 'functions.do'

const results = await ai.writeBlogPostTitles({ 
  topic: 'automating business workflows with LLMs',
  audience: 'executives',
  count: 10
})

console.log(results)
// {
//   blogPostTitles: [
//     '10 Steps to Automate Your Business Workflows with LLMs',
//     'The Executive's Guide to LLM-Powered Process Automation',
//     'Transforming Executive Decision-Making with AI Workflows',
//     'How LLMs Are Revolutionizing C-Suite Productivity',
//     'Strategic Implementation of LLMs: An Executive Roadmap',
//     'Competitive Advantage Through Intelligent Automation',
//     'Measuring ROI on LLM Integration in Enterprise Workflows',
//     'From Boardroom to Automation: The Executive's LLM Playbook',
//     'Scaling Business Operations with AI-Driven Workflow Solutions',
//     'Future-Proofing Your Business with Intelligent LLM Systems'
//   ]
// }
```

Alternatively, you can define the return type of the function during initialization to get strongly-typed responses:

```typescript
import { AI } from 'functions.do'

const ai = AI({

  createLeanCanvas: {
      problems: ['top 3 problems the business solves'],
      customerSegments: ['target customers and users for the product'],
      uniqueValueProposition: 'clear and compelling message that states why you are different and worth buying',
      solutions: ['outline of the solutions to the identified problems'],
      unfairAdvantage: 'something that cannot be easily copied or bought',
      revenueStreams: ['revenue model, lifetime value, revenue, gross margin'],
      costStructure: ['customer acquisition costs', 'distribution costs', 'hosting', 'people', 'etc.'],
      keyMetrics: ['key activities you measure (acquisition, retention, referrals, etc.)'],
      channels: ['path to customers (inbound, outbound, viral, etc.)'],
      earlyAdopters: 'characteristics of the ideal early adopter'
    }

})


const results = await ai.createLeanCanvas({ domain: 'aws.amazon.com' })
```


Finally, you can also define the return type and override system settings for a specific function like:

```typescript
import { ai } from 'functions.do'

const results = await ai.generateLandingPage({ 
  brand: 'Functions.do', 
  idea: 'AI-powered Functions-as-a-Service' 
}, {
  model: 'anthropic/claude-3.7-sonnet:thinking',
  system: 'You an expert at generating highly-converting marketing copy for startup landing pages',
  temperature: 1.0,
  seed: 1741452228,
  schema: {
    headline: 'attention-grabbing headline that clearly states value proposition',
    subheadline: 'supporting statement that adds clarity to the headline',
    productDescription: 'concise explanation of what the product does and its benefits',
    keyFeatures: ['list of main features or benefits'],
    socialProof: ['testimonials, user counts, or other trust indicators'],
    captureForm: 'description of email capture form and call to action',
    incentive: 'what users get for joining the waitlist (early access, discount, etc.)',
    visualElements: 'description of images, videos, or other visual elements',
    faq: [{
      question: 'frequently asked question related to an objection',
      answer: 'clear answer to the question that overcomes the objection'
    }],
    launchDate: 'expected product launch date or timeline',
    callToAction: 'primary button text and action',
    secondaryAction: 'secondary button text and action'
  }
})