Rag Conversation Composer

Scale your expertise by packaging it into shareable AI conversations with custom context, guidance, and intelligence.

Turn any consultation, support interaction, or knowledge transfer into a ready-to-use AI conversation that your clients, customers, or team members can use instantly—without you being there.

---

Why Use This Actor?

The Problem

You're spending hours answering the same questions, providing consultations, or sharing expertise one-on-one. Your knowledge doesn't scale, and your time is limited.

The Solution

Package your expertise once into an AI conversation that includes:

• Your guidance (custom instructions that make the AI think like you)
• Relevant context (customer data, documentation, background info)
• Smart recommendations (suggested tools and resources)

Recipients get a copy-paste prompt they can use in Claude, ChatGPT, or Gemini—with all your knowledge and context pre-loaded.

The Result

• 10x your reach: One package = unlimited conversations
• 24/7 availability: Your expertise works while you sleep
• Consistent quality: Every interaction follows your guidance
• Zero marginal cost: No extra work per recipient

---

Real-World Use Cases

🎯 Customer Success

Problem: Customers need personalized guidance, but you can't scale 1:1 consultations.

Solution: Create packages with their usage data and optimization recommendations.

Example: "Based on your $2,000/month Instagram scraping usage, here are 3 ways to reduce costs by 30%..."

Impact: Handle 100+ customer consultations simultaneously.

---

💼 Sales & Pre-Sales

Problem: Prospects need custom demos and technical answers before buying.

Solution: Generate proposal packages with their requirements and solution fit analysis.

Example: "For your healthcare compliance needs with 200 users, here's how our solution compares to your current setup..."

Impact: Qualify and educate prospects without sales engineering time.

---

📚 Customer Education

Problem: Customers struggle with onboarding and need hand-holding.

Solution: Create guided learning experiences with their specific use case.

Example: "Let's set up web scraping for your e-commerce analytics—I have your API keys and data requirements ready..."

Impact: Self-service onboarding with 10x better completion rates.

---

🤝 Team Knowledge Sharing

Problem: Important context lives in people's heads, slowing down collaboration.

Solution: Package project context for cross-team handoffs.

Example: "Here's everything about the Q4 migration project: architecture decisions, tech stack, pending issues..."

Impact: Instant context transfer, zero ramp-up time.

---

How It Works (3 Simple Steps)

1️⃣ You Create the Package

Run the Actor with:

• Title: "Q1 Business Review - Acme Corp"
• Instructions: "You're a business analyst. Analyze usage and suggest growth opportunities."
• Context: Customer's usage data, benchmarks, recent tickets
• Optional: Recommended MCP servers for deeper functionality

Time required: 2-5 minutes

---

2️⃣ Actor Generates the Package

The Actor creates a beautiful markdown file containing:

• Setup instructions for the recipient
• A ready-to-paste AI prompt with all context embedded
• MCP recommendations
• Conversation starter

Processing time: 5-30 seconds

---

3️⃣ Recipients Use It Instantly

They:

1. Open the markdown file
2. Copy the "Ready-to-Use Prompt"
3. Paste into Claude/ChatGPT/Gemini
4. Start chatting with full context

No signup, no installation, no complexity.

---

Quick Start

Minimal Input (Just 2 Fields Required!)

{
  "conversationTitle": "Customer Support - Help with API Integration",
  "guidanceInstructions": "You're a helpful API support engineer. Guide users through integration step-by-step."
}

Full-Featured Example

{
  "conversationTitle": "Enterprise Upgrade Consultation",
  "recipientName": "Sarah Johnson",
  "guidanceInstructions": "You're a customer success manager at Apify. Help customers understand Enterprise plan benefits based on their usage. Be consultative, not pushy.",
  "ragContext": [
    {
      "title": "Customer Usage",
      "content": "Monthly Spend: $2,000\nPrimary Actor: Instagram Scraper\nTeam Size: 5 users",
      "source": "Apify API"
    },
    {
      "title": "Enterprise Benefits",
      "content": "Priority support, SLA guarantees, 15-30% volume discounts, custom development"
    }
  ],
  "recommendedMCPs": [
    {
      "name": "Apify MCP",
      "url": "https://mcp.apify.com/?tools=docs,actors",
      "description": "Access Apify docs and Actor details during conversation"
    }
  ],
  "conversationStarter": "I'd love to understand how Enterprise plan could benefit your Instagram scraping workflow.",
  "outputFormat": "markdown"
}

---

Input Reference

Required Fields

Field	Type	Description
conversationTitle	string	Descriptive title (e.g., "Q4 Review - Acme Corp")
guidanceInstructions	string	How the AI should behave (your expertise)

Optional Fields

Field	Type	Description
recipientName	string	Who receives this package
ragContext	array	Knowledge documents with title, content, source
webDataSources	array	URLs to auto-fetch and include as context
recommendedMCPs	array	MCP servers with name, url, description
conversationStarter	string	Initial prompt to kick off conversation
outputFormat	enum	markdown, json, or both (default: markdown)

---

How to Add Context Data

There are multiple ways to populate the RAG Context Data field:

Option 1: Manual JSON Input

Directly paste JSON into the field (best for testing):

[
  {
    "title": "Customer Profile",
    "content": "Company: Acme Corp\nIndustry: E-commerce\nTeam Size: 50",
    "source": "CRM"
  }
]

Option 2: Fetch from URLs

Use the webDataSources field to automatically fetch content:

{
  "webDataSources": [
    "https://api.example.com/customer/123",
    "https://docs.example.com/knowledge-base"
  ]
}

The Actor will fetch and include this content automatically.

Option 3: From Apify Dataset

If you have data from another Actor:

1. Run an Actor that outputs to a Dataset
2. Copy relevant data from the Dataset
3. Format as ragContext JSON array

Option 4: API Integration (Production)

Call this Actor programmatically with context from your system:

const client = new ApifyClient({ token: 'YOUR_TOKEN' });

await client.actor('username/rag-conversation-composer').call({
  conversationTitle: "Consultation - Customer Name",
  guidanceInstructions: "...",
  ragContext: [
    {
      title: "Customer Data",
      content: JSON.stringify(customerData),
      source: "Database"
    }
  ]
});

Option 5: Upload JSON File

If the Apify UI supports it, you can upload a prepared JSON file with your context data.

---

Output & Usage

What You Get

• Markdown file: Beautiful, shareable conversation package
• JSON file (optional): Machine-readable format for integrations
• Dataset entry: Metadata with URLs and statistics

How Recipients Use It

The markdown file contains everything they need:

# Ready-to-Use Prompt

Copy this entire section and paste into Claude, ChatGPT, or Gemini:

[Full prompt with your instructions + all context embedded]

That's it. No API keys, no setup, no learning curve.

---

Key Benefits

⚡ Instant Scalability

Create once, use unlimited times. Your expertise becomes infinitely scalable.

🎯 Personalized at Scale

Each package includes recipient-specific data and context—feels 1:1 but scales to thousands.

🔒 Privacy First

• All data stays in your Apify storage
• Recipients only see what you include
• No tracking, no external calls (except optional web fetches)

🌐 Platform Agnostic

Works with Claude, ChatGPT, Gemini—recipients use their preferred AI.

💰 Zero Marginal Cost

No API costs per conversation—recipients use their own AI accounts.

---

Advanced Features

Auto-Fetch Web Content

{
  "webDataSources": [
    "https://docs.example.com/api",
    "https://example.com/customer-profile"
  ]
}

Actor fetches and includes this content automatically.

MCP Integration

Recommend Model Context Protocol servers for enhanced capabilities:

{
  "recommendedMCPs": [
    {
      "name": "Apify MCP",
      "url": "https://mcp.apify.com/?tools=docs",
      "description": "Real-time access to Apify documentation"
    }
  ]
}

Multiple Output Formats

• markdown: Human-readable (default)
• json: Machine-readable for integrations
• both: Get both formats

---

Technical Details

• Runtime: Node.js 20+
• Dependencies: Apify SDK, Axios, Cheerio
• Storage: Key-Value Store (files) + Dataset (metadata)
• Performance: 5-30 seconds per package
• Limits: No hard limits on context size (reasonable usage recommended)

---

Privacy & Security

✅ All data stored in your Apify account ✅ No external API calls (except optional URL fetches) ✅ Recipients see only what you include ✅ No tracking or analytics ✅ No secrets or API keys required

---

Support & Resources

• Documentation: You're reading it!
• Issues: GitHub Repository
• Apify Discord: Join Community
• Apify Docs: Platform Documentation

---

License

Apache-2.0

---

Built with ❤️ using the Apify platform

Turn your expertise into scalable AI conversations. Start creating packages today.