← Back to Home

SoyManada — Verified Directory for Hispanic Immigrants

soymanada.com2025Founder & Developer

Context & Problem

Canada welcomes over 400,000 new permanent residents per year — a growing share from Latin America. Spanish-speaking immigrants navigating the IRCC process face a critical gap: finding trustworthy, verified service providers in their own language. Directories are fragmented, English-only, and offer no verification layer — leaving newcomers vulnerable to unqualified or fraudulent consultants.

  • No centralized, language-aware directory for Hispanic migrants.
  • Zero verification standards for listed service providers.
  • Reviews impossible to trust without a minimum volume threshold.
  • No FR-CA coverage — ignoring Canada's bilingual market reality.
  • No PIPEDA-compliant data handling in existing informal directories.

My Role: Founder & Full-Stack Developer

I designed and built this product from zero — covering the full value chain: business strategy, product definition, UX design, and complete technical implementation. No team, no outsourcing. Every architectural decision, from data model to deployment pipeline, was made and executed by me.

The Solution

  • Freemium model with 3 tiers: Bronze (free) / Silver ($10 USD/mo) / Gold ($20 USD/mo) — designed for progressive monetization without friction for providers starting out.
  • Manual verification pipeline: Every provider is reviewed through a custom AdminPanel before going live. No automated approval — trust is the product.
  • Real multilingual routing (ES / EN / FR-CA): Not just translated text — routes, content, and metadata are fully localized for each market. FR-CA is a strategic differentiator for Canada's bilingual environment.
  • Review system with visibility threshold: A provider's reviews only go public after a minimum of 3 are collected — preventing manipulation and ensuring signal quality.
  • Native application form: Providers apply directly through the platform with a form that pipelines data straight to the database — no manual intake process.
  • PIPEDA-aligned data handling: Minimal personal data collection, explicit consent flows, and user-controlled data — compliant with Canada's federal privacy law from day one.

Technical Architecture

Frontend
React + Vite
i18next for multilingual routing
Backend / DB
Supabase
PostgreSQL + RLS + Auth
CI/CD
GitHub Actions
Auto-deploy on push to main

Key technical decisions:

  • RLS policies in Supabase to cleanly separate public data (visible to all), private data (providers only), and admin data — no middleware needed to enforce access control.
  • Dynamic type migration (UUID → TEXT) in PostgreSQL executed without downtime — maintaining data integrity across a live system.
  • Custom AdminPanel built in-house to manage provider verification, subscription tiers, and review visibility — no third-party admin dependency.
  • PIPEDA compliance by design — data minimization, consent-first flows, and no third-party data sharing without explicit user authorization. Built for the Canadian regulatory standard, not retrofitted to it.

Strategic Impact

SoyManada directly addresses a real gap in Canada's immigrant services ecosystem. By building with FR-CA support from day one, it operates natively within Canada's bilingual market — a rare differentiator among Spanish-language platforms.

"This project reflects my end-to-end product thinking: I identified the gap, validated the model, and shipped a fully functional SaaS — alone, from scratch, while working full-time."

Development Methodology

Shipped as a solo founder using a structured multi-LLM development pipeline — allowing one person to operate at the research depth and iteration speed of a small team.

Research
Perplexity
Market validation, competitive landscape, immigration ecosystem analysis
Architecture
Claude
System design, data model decisions, RLS strategy, code review
Prototyping
ChatGPT
Rapid UI iteration, copy generation, edge case exploration
Validation
Gemini
Cross-validation of decisions, alternative approaches, blind spot checks

Each model serves a specific role in the pipeline. The decisions, product ownership, and execution remain entirely human — AI is the toolchain, not the author.

Project Status: Live — soymanada.com
Visit soymanada.com Discuss this Project