Winners of the Apify MCP server configuration competition

It’s time to announce the winners of the Apify MCP server competition. Over the past month, developers and vibe coders have explored the capabilities of the newly built Apify MCP Configurator, and submitted their workflows for the contest. We’d like to thank everybody who participated.

Contestants used the MCP Configurator to select and combine Actors from Apify Store, design system prompts, and demonstrate real-world use cases through their chosen MCP clients (Claude Desktop, Cursor, or the Apify Tester Client).

Each submission was evaluated through four criteria:

Functionality (25 points): Practical utility and appropriate Actor selection
System prompt quality (25 points): Effectiveness in guiding AI behavior
Demonstration (25 points): Quality of example prompts
Use case value (25 points): Real-world application and innovation

The winners

First place: Aditya Akuskar - AI financial analyst

Aditya took home the first prize with a comprehensive AI financial analyst configuration. Using Claude Desktop as his client, he assembled a toolkit of eight specialized financial Actors that work together:

Multiple Yahoo Finance scrapers for stock data
A crypto price history API for cryptocurrency analysis
AI-powered financial analyst Actors
RAG web browser for contextual financial research

What set Aditya's submission apart:

Aditya’s example prompts showcased versatility. They analyzed Microsoft's quarterly revenue trends and balance sheets, but could also track Bitcoin price patterns over a 90-day period.

The configuration addresses the needs of investors, traders, and financial analysts who require access to reliable market data and quick analysis. Each Actor serves a clear purpose in the overall workflow, demonstrating a deep understanding of financial analysis requirements.

Use case: Stock market analysis, financial data scraping, cryptocurrency tracking

Try Aditya's configuration

MCP server URL:

<https://mcp.apify.com/?tools=actors,runs,commemorative_amethyst/yahoo-finance-scraper,websift/crypto-pricehistory-api,bala-ceg/ai-financial-analyst,pintostudio/stock-information-investing-com,architjn/yahoo-finance,katzino/market-mind-ai,eraydiler/yahoo-finance-historical-data-scraper,apify/rag-web-browser>

Example test prompts:

"Get 6 months of historical price data for Microsoft (MSFT), then retrieve its latest income statement and balance sheet. Compare the quarterly revenue trends and show me the latest analyst recommendations."
"Get Bitcoin (BTC) price history for the last 90 days using the crypto price history Actor. Analyze the trend and identify any significant price movements or patterns."
"Find an Actor to scrape product data from Amazon, run it for 'wireless headphones,' then retrieve the dataset and show me only the product names, prices, and ratings sorted by price."

System prompt highlights: Aditya configured his agent as an AI financial analyst with access to real-time financial data through Apify Actors. The system prompt clearly defines capabilities, including fetching stock data, analyst recommendations, and cryptocurrency price history. It also features structured output formatting with clear headings, key metrics, and professional financial insights.

Second place: Pratheek Raj Urs C P - AI research intelligence agent

Pratheek secured second place with a refined AI research intelligence agent built for Cursor. What made his submission noteworthy was the evolution we observed: he submitted two versions, a comprehensive v1 and a streamlined v2. The simplified version actually scored higher.

Configuration highlights:

Six carefully curated research Actors (RAG browser, web scrapers, search, and crawlers)
Concise system prompt (v2’s 800 words vs. v1's 4,500)
Balance of functionality and token efficiency

Pratheek’s test prompts were realistic and executable: researching AI-powered customer service tools and conducting due diligence on companies. This practical approach demonstrated that he had built something that could be deployed in production immediately.

Use case: Market research, competitive intelligence, company due diligence

Try Pratheek's configuration

MCP server URL:

<https://mcp.apify.com/?tools=actors,docs,experimental,runs,storage,apify/rag-web-browser,apify/web-scraper,apify/website-content-crawler,apify/cheerio-scraper,apify/google-search-scraper,peterasorensen/snacci>

Example test prompts:

"Research the current state of AI-powered customer service tools. I need to understand the key players, their pricing models, recent funding rounds, and customer sentiment. Focus on companies that raised funding in 2024-2025."
"Conduct due diligence research on 'GreenTech Solutions Inc.,' a fictional sustainable energy startup. I need their business model, leadership team, recent news, social media presence, and any red flags."

System prompt highlights:

Pratheek's system prompt positions the agent as an "AI research intelligence agent powered by Apify's web scraping and data collection capabilities." It defines a clear 5-step workflow (query analysis - multi-source gathering - cross-validation - synthesis - structured output) and specifies core capabilities including web search, content extraction, social media intelligence, and data synthesis. The output format is concise: executive summary, detailed analysis, key insights, and sources.

Third place: Pedro Farias - Instagram content creator assistant

Pedro earned third place with a focused, niche-specific solution: an Instagram content creator assistant which uses the Apify Tester Client. While other participants built broad research tools, Pedro identified a specific high-value market and developed a focused solution for it.

Configuration details:

4 Instagram-specific Actors (scraper, reel scraper, post scraper, profile scraper)
Portuguese language support demonstrates international appeal
Automated workflow from Instagram scraping to ready-to-publish content

Why it stood out:

Pedro's submission exemplified the principle that doing one thing well often beats trying to do everything adequately. His configuration addresses a real pain point for social media managers and content creators: the time-consuming process of content research and ideation.

The test prompt (in Portuguese) demonstrated scraping recent Instagram posts and generating social media-ready content complete with engaging captions, emojis, and relevant hashtags. The result is workflow automation that saves content creators hours of manual work.

Use Case: Social media content generation, Instagram analysis, digital marketing

Try Pedro's configuration

MCP server URL:

<https://mcp.apify.com/?tools=actors,docs,experimental,runs,storage,apify/instagram-scraper,apify/instagram-reel-scraper,apify/instagram-post-scraper,apify/instagram-profile-scraper>

Example test prompt: "Raspe os últimos 5 posts do perfil @openai no Instagram e gere um resumo pronto para postagem em redes sociais com legenda atraente, emojis e hashtags relevantes."

(Translation: "Scrape the last 5 posts from the @openai Instagram profile and generate a summary ready for social media posting with attractive captions, emojis, and relevant hashtags.")

System prompt highlights: Pedro's system prompt (in Portuguese) positions the agent as a specialist in creating engaging, educational, and fun content for Instagram and Facebook. It emphasizes creating posts that appear human-written, with strategic use of emojis, relevant hashtags, and clear calls to action. The agent is also configured to generate visual prompts for 1080x1080 marketing images that complement the social media posts. It delivers three caption variations per theme plus corresponding image generation prompts.

Notable mentions: other competitors

While only three could win prizes, several other submissions demonstrated creativity and technical skill. We’ve selected a few of them.

Muhammet Akkurt - ChainSight On-Chain Intelligence Analyst

Muhammet created the competition's most original submission with a specialized blockchain intelligence analyst for the Solana network. What made this exceptional was that all five Actors were custom-built by Muhammet himself specifically for this use case:

gmgn-copytrade-wallet-scraper, which discovers profitable wallets
gmgn-wallet-stat-scraper, for performance statistics
gmgn-wallet-activity-scraper, shows recent transactions
arkham-intelligence-wallet-data-scraper, scrapes identity data
arkham-ai-transaction-analyzer, provides AI analysis

His submission scored well (90/100) and showed the most technical ambition. The custom Actors demonstrated his determination in solving a specific problem. For crypto traders seeking alpha signals and copy-trading opportunities, this configuration provides a competitive advantage.

Client: Gemini CLI

Standout Quality: Only submission with an entirely custom-built Actor ecosystem

Luděk built a viral content analyzer for LinkedIn using Claude Desktop. His configuration focused on identifying trending content and engagement patterns across LinkedIn posts about AI, LLM, RAG, and generative AI topics.

The submission featured a detailed system prompt (over 5,000 words) that allowed the following:

Engagement metrics classification (viral, trending, popular)
Viral content pattern analysis
Author authority impact assessment
Interactive analytics dashboard generation

Client: Claude Desktop

Actors: Two LinkedIn scrapers (no-cookies scraper, profile posts)

Standout Quality: Most detailed analytical methodology and dashboard visualization

Gustavo Andres Liberto Salcedo - due diligence analyst

Gustavo submitted a company investment due diligence configuration with a clear, professional use case that should resonate with investors and VCs. His system prompt outlined a solid research protocol covering corporate information, funding history, team analysis, and red flag identification.

While his submission was limited by having only one Actor configured (Website Content Crawler), the use case itself was valuable. We felt it represented a missed opportunity: with proper Actor selection (search, social media, news scrapers), this could have been a contender for a winning spot.

Client: Gemini CLI

Key takeaway: Even the best use case needs adequate tooling to execute effectively

Best practices from the competition

Focused Actor selection - Winners chose 4-8 Actors that worked together coherently, avoiding both over-engineering (15+ Actors) and under-tooling (1 Actor) so the optimal configuration is 5-7 carefully selected Actors.
Quality over quantity - The best submissions had concise, effective system prompts (800-1,200 words) rather than overwhelming detail. It's always good to test with 2-4 diverse test prompts.
Practical demonstration - Top entries showed realistic, executable use cases rather than overly ambitious scenarios targeted toward a specific target audience
Clear value proposition - Winners solved problems for specific target audiences (traders, researchers, content creators)

Common pitfalls to avoid:

Feature bloat with too many Actors
Overly complex system prompts that waste tokens
Mismatched capabilities (promising what tools can't deliver)
Insufficient demonstration with only one test prompt

Try it yourself

Inspired by these winning submissions? Create your own MCP server configuration at mcp.apify.com. The Apify MCP Configurator provides the building blocks to bring your AI agent ideas to life.