Video summary

Office Hours with Matthew & Flo

Main summary

Key takeaways

Gaming

Storyline / Context

  • The video is an “Office Hours” session with Matthew and Flo, focused on building a World Cup-themed betting app using WeWeb and an MCP server integration.
  • A major theme is how the team uses real data (World Cup matches) and AI-assisted tooling to scaffold both the frontend and backend/payment-related logic—culminating in a near-working demo of placing and settling bets.

“Gameplay” / Build Highlights (development workflow + demo)

Even though this isn’t a traditional game, the “gameplay” is the development workflow and demo build:

World Cup-themed prediction/betting app concept

They scaffold core app functionality, including:

  • Data modeling for tables and match data
  • Authentication
  • UI screens such as:
    • Login / Home
    • Matches
    • Admin pages
    • Leaderboard
  • User flow:
    • Users can place bets
    • Bets can later be settled based on match outcomes

AI + MCP server workflow

They demonstrate an AI workflow using an MCP tool access prompt (described as “popup” guidance) so the AI can retrieve project context and relevant information (e.g., World Cup info), then generate implementation details.

The AI is used to build things like:

  • A model for betting predictions (using a real money concept, but implemented with fake currency/odds for demo constraints)
  • Features such as:
    • Exact score predictions
    • User accounts
    • Match documentation-driven endpoints

Live data and iterative debugging

The demo progresses by progressively wiring:

  • Teams/matches data
  • UI components and navigation
  • Authentication and access checks
  • Betting logic and settlement flows

Some parts were not fully wired yet, especially:

  • Payments integration
  • At least one rules/permission validation issue related to settlement

Key Strategies / Tips Mentioned

Use AI with MCP for endpoint generation

  • Provide the AI a documentation URL (from payment providers or other APIs).
  • The AI can help generate the required endpoints, then you manually test and adjust.

Demo/scaling readiness

The MCP server is described as rolling out gradually:

  • Early access / continued testing during the current week
  • Likely full go-live next week, depending on scaling and fixes

Payments integration caveat

  • Stripe integration is not yet available via the MCP server.
  • Payments features may require:
    • Non-MCP “proper actions”
    • or backend-side work
    • until MCP payment connectors mature

Protect apps from bots / DoS

  • Monitor request spikes via usage dashboards.
  • Apply backend protections/middleware and rate-limiting-style approaches.
  • If attacks occur, contact support for help; they consider native rate limiting a future improvement.

Regulatory / data residency (Kazakhstan) approach

To meet local storage laws (Kazakhstan):

  • Self-host the backend with local databases
  • Ensure EU vs Kazakhstan placement is handled carefully
  • Keep personal-data tables in the required region
  • Alternatively:
    • store only the personal DB locally
    • expose only the needed data via endpoints to external services

WeWeb vs Zano backend choice

General guidance:

  • Prefer WeWeb for:
    • full-stack needs with structured requirements
    • quick self-hosting
    • smaller backends
  • Prefer Zano when you need:
    • more complex backend structure (e.g., agents)
    • multiple instances/branches/calculations
    • corporate/infrastructure-heavy setups

They emphasize this is not framed as direct competition—frontend-only and full-stack/backend approaches both have valid use cases.

Featured Gamers / Sources (mentioned)

  • No specific esports teams or publishers are featured.

People/roles mentioned

  • Matthew
  • Flo
  • Audience/users by handle:
    • AD Gamer
    • Muriel
    • Nikki
  • “Aaron Aron” is referenced as asking a question

Payment/banking reference

  • Hali Bank (Kazakhstan bank)

Tools/platforms discussed

  • Claude / Claude Code
  • ChatGPT (mentioned in AI tooling context)
  • WeWeb
  • Zano
  • Stripe (not yet available via MCP server)
  • Supabase/Superbase (mentioned in passing in their setup context)
  • MCP servers and integrations
  • Visual Studio / Visual Studio integration and VS Code (in the context of using MCP + Zano together)

Original video