Vibe Coding Showdown:
which build bot actually ships?
We ran the same prompt across Bolt, v0, Lovable, Replit and Raq.com. Only one returned a working solution.
Tested with identical prompt
The Test: UK Company Financial Research Tool
We asked each platform to build the same complex internal tool with the exact same prompt
The Full Prompt Given to All Platforms
"I want to create the ultimate UK company financial research tool for my team ✨ How it should work: 🏢 User enters minimum 2 UK company names via autocomplete dropdown (pulls from Companies House API) 🔍 Auto-validates that financial data exists with FinUK API (no ghost companies!) ⚡ Kicks off background processing to pull all data from Companies House and FinUK APIs 🤖 Gemini 2.5 Pro performs deep comparative analysis across ALL companies - spotting trends, benchmarking performance, and highlighting competitive advantages 👥 Everyone on the team can access all reports 📈 Include nice formatting of the original API data 🎯 User flow: Search and select company names from dropdown → system validates → background processing starts ✨ Only ask for company names - we don't need title or description User can navigate away and continue working whilst reports generate Come back to find professional-grade analysis waiting with nice formatting 💪 Team browses reports in clean list view with search/filters Click any report to see nicely formatted API data (tables) + formatted AI comparative insights 📋 In summary – what the tool delivers: ✅ Smart company search: Autocomplete powered by Companies House API, ensuring only valid UK companies are selectable ✅ Instant validation: Automatic checks via FinUK API so only companies with real, usable financial data are processed ✅ Deep AI insights: Gemini 2.5 Pro runs comparative analysis across all selected companies—highlighting trends, benchmarks, and competitive advantages ✅ Beautifully formatted data: - 📊 Companies House data tables – clean, structured, and easy to review - 💷 FinUK financial tables – detailed and neatly presented - ✨ AI‑generated reports – professional‑grade summaries with clear visual formatting ✅ Team‑ready access: Reports are stored, searchable, and filterable 🔌 APIs: 🏛️ Companies House API: API Key: 61a67af5-3ae1-4cd0-b02b-a400758b094d API Docs: https://developer-specs.company-information.service.gov.uk/companies-house-public-data-api/reference Example Companies: Milton Homes Limited, Ecotrade Europe Ltd, Flavour Warehouse Limited 💰 FinUK API: API Key: finuk_6b07449efde592c0556ec825378b373ec4903b95fc879cc8064ac4d5a65da84f API Docs: https://finuk.io/docs 🧠 OpenRouter API: API Key: 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 API Docs: https://openrouter.ai/docs/quickstart ⚠️ You must build everything in one go – no back and forth"
What We Asked Each Platform to Build
A complete UK company financial research tool with these specific requirements:
Smart Company Search
Autocomplete dropdown powered by Companies House API with automatic validation via FinUK API to ensure only real companies with financial data
Background Processing
Asynchronous data pull from multiple APIs while users can navigate away and continue working—no need to keep the tab open
AI Comparative Analysis
Gemini 2.5 Pro performs deep analysis across all selected companies, spotting trends and highlighting competitive advantages
Team-Ready Access
All team members can access generated reports with search and filtering capabilities in a clean, professional interface
Professional Formatting
Clean data tables for Companies House and FinUK data, plus beautifully formatted AI-generated reports with visual hierarchy
Production Ready
Must work end-to-end with real API integrations, proper error handling, and database storage—not just a prototype
The Results
Watch how each platform handled the same complex task - videos are sped up by 24x during generation
Bolt.new
❌ Non-working
Vercel v0
❌ Non-working
Lovable.dev
❌ Non-working
Replit
❌ Non-working
Raq.com
✅ Working Solution
Bolt.new | v0 | Lovable | Replit | ||
|---|---|---|---|---|---|
| Continues if closed | Yes | No | No | No | No |
| Database set up | Automatic | Not in this case | Manual Supabase | Manual Supabase | Automatic |
| Auto-fixing issues | Automatic | No | No | No | Attempted |
| Final deliverable | Working app | Errors | Errors | Errors | Not functioning |
Why Did Raq.com Work?
While others failed, Raq.com delivered a complete, working solution.
Raq.com
Self-Correcting AI
Checks itself and fixes issues automatically without user intervention
Researches APIs thoroughly to understand real data structures
Makes mistakes but fixes them autonomously during development
Delivers working solutions that actually function end-to-end
Others
Limited Self-Correction
Don't do enough checking of their own implementations
Hallucinates/guesses API structures without proper research
Expect users to debug technical issues manually
Difficult for non-technical users to get working solutions without continuous manual error looping
The Key Difference
Raq.com's AI agent continuously validates and improves its work, while other platforms generate code and hope it works. Raq.com brings the power of Claude Code agentic AI to non-technical users.
Experience the Difference
Get early access to the only platform that actually delivers working internal tools for non-technical users.