⚽ Global Football Companion App: Feel the World Cup, Don't Just Watch It
An idea for the AWS Builder Center Weekend Challenge — and why this one is worth your time even if you can't enter
The Weekend Challenge this month lands at exactly the right moment: with the world's biggest football tournament kicking off this summer, Ben Fowler asks what companion app you'd build for fans and a genuinely good question on the table.
I can't enter formally — the terms and conditions don't allow it. But the question was too interesting to let pass without an answer. So instead of a comment, I wrote this.
The Idea: Global Football Companion App
Most football companion apps — FotMob, Sofascore, LiveScore — do a great job solving the score problem. They tell you what happened, who scored, where the standings sit. Fast and well-covered.
But there's a problem every one of them leaves open: what if you can't look at the screen?
I'm not talking about distraction or multitasking. I'm talking about the 2.2 billion people worldwide who live with some form of visual impairment, according to the WHO. For many of them, following a match in real time still means depending on someone standing next to them narrating the action, or a radio broadcast that isn't always available in the right language or with the level of detail they actually want.
Global Football Companion App is built to close that gap.
What It Does
The app takes structured match events — goal, shot, card, substitution, kickoff, penalty — and converts them into three simultaneous outputs:
- Generated audio: a short, clear phrase. Not improvised commentary. "Minute 34. Goal for Argentina. Long-range shot, unmarked."
- Encoded haptic patterns: a different vibration signature for each event type. A goal feels different from a yellow card. A dangerous attack feels different from a lineup change.
- Structured WebVTT captions: a text history of the match — accessible, exportable, and useful for replay.
The core idea is that the event isn't just described — it's felt. A goal isn't only a sentence. It's a tangible signature you can recognize without looking at the screen, without waiting for someone to tell you what happened.
Who It's For
The primary audience is blind and low-vision football fans who today have no app genuinely designed for them. Not one where accessibility is a checkbox — one where the non-visual experience is the product, from day one.
But there are other use cases that aren't obvious at first:
- People following matches in transit, phone in their pocket.
- Fans in noisy environments where the broadcast audio gets lost.
- Older users who prefer less visual density and more audio clarity.
- Supporters learning the tournament's language who need adapted narration.
This is the classic pattern of inclusive design: when you build seriously for the most constrained case, you end up building something better for everyone.
The Feature That Sets It Apart
Not the score. Not the stats. The encoded sensory layer.
The difference between "I'll notify you there was a goal" and "you can feel there was a goal" sounds small. It isn't. The second option shifts the experience from depending on an external channel toward a controllable, repeatable, and personalizable interface.
The three modes the app would offer — all fed from the same event model, the user picks their style:
| Mode | For whom | What it delivers |
|---|---|---|
| Narration Mode | Fans who want continuous commentary | Short audio per event, on-screen text, full history |
| Pulse Mode | Fans who want only the critical moments | Vibration + alert for goals, cards, and final whistle only |
| Tactical Mode | Fans who want more data density | Possession, offensive patterns, performance comparisons |
How It's Built on AWS Free Tier
This is the part I found most interesting to work through.
AWS overhauled its Free Tier model on July 15, 2025. Accounts created after that date receive up to $200 in credits plus a set of Always Free services that never expire. The key Always Free services for this architecture — Lambda, DynamoDB, and Cognito — are permanent regardless of account age. Other services like API Gateway and Amplify Hosting draw from the credit pool for newer accounts. With that in mind, the proposed architecture is deliberately lean:
The key services and why they fit:
| Service | Role | Free limit | Type |
|---|---|---|---|
| AWS Amplify Hosting | Accessible PWA frontend + CDN | 1,000 build min/mo · 15 GB transfer | Credits / Always Free UI |
| Amazon Cognito | Identity and user preferences | 10,000 MAU/month | Always Free |
| API Gateway | HTTP API between Lambda and client | 1M HTTP calls/month | Credits (new accounts) |
| AWS Lambda | Match event normalization | 1M requests + 400K GB-seconds/mo | Always Free |
| Amazon DynamoDB | Match state cache and preferences | 25 GB storage · 200M requests/mo | Always Free |
| EventBridge Scheduler | Scheduled fixture and lineup refresh | 14M invocations/month | Always Free |
| Amazon Polly | Cloud voice (demos or fallback only) | 5M characters/month Standard | Credits |
The most important design decision: the MVP defaults to on-device TTS rather than Polly as the primary voice route. That brings cost to near zero and eliminates the risk of exhausting credits under concurrent load. Polly stays as an optional fallback for more natural-sounding voices in demos.
Open and Free Data Sources
The data layer is a combination of sources, each with a specific role — and honest about what each one actually provides for free:
- API-Football (by API-Sports) — the primary source for live events: goals, cards, substitutions, lineups, and real-time fixture status. Free tier: 100 requests/day, includes all core endpoints. Sufficient for a caching-first MVP with smart polling intervals.
- football-data.org — fixtures, standings, and competition metadata for the top 12 competitions. Free forever, but the free tier does not include lineups or player-level events — those require a paid plan. Used here for schedule context and competition structure only.
- TheSportsDB — metadata enrichment, crests, TV listings. Free tier available, good complement for visual assets and competition context.
- OpenLigaDB — open JSON API, no authentication required, ODbL license. Useful for prototyping and offline demos without any quota pressure.
- StatsBomb Open Data — historical event data on GitHub. Not for live use, but invaluable for designing narration templates and testing event vocabulary before connecting to a live feed.
One important note across all of these: there is no free, public FIFA API for minute-by-minute match events. The official World Cup data is licensed commercially via Stats Perform. The MVP needs to live on documented open sources — which, with smart caching, is sufficient for an accessible second-screen product.
What Would Remain Open in a Real MVP
Being honest about the limitations:
- Haptics on the web are Android-only in practice. The web Vibration API works on Android Chrome but Safari has never supported it, WebKit has published a formal position opposing the spec, and Firefox removed support in version 129. For iOS users, haptics requires a native app — not just "better control," it's the only path. A PWA is the right starting point for prototyping; native is required for a real iOS experience.
- 100 requests/day on API-Football's free tier is tight for a live match with polling every 30 seconds. Smart caching in DynamoDB and differential updates in Lambda are not optional — they're what makes the free tier work.
- The real latency of free APIs under a major match (World Cup final, millions of concurrent users globally) has to be measured, not assumed. Free-tier SLAs are not production SLAs.
- Co-design with actual blind users from day one — not after building. Preferably with organizations that already work in sports accessibility, like those behind projects such as Escoita (Real Sociedad's in-stadium audio description service).
Why This Challenge Is Worth Your Attention
The Weekend Challenge asks for a companion app. Most responses will probably lean toward ML prediction engines, stats dashboards, or Q&A bots. All of that is valid.
But the most interesting opportunity, I think, is the space that every existing app leaves completely empty: an experience designed from scratch for people who can't — or don't want to — depend on their eyes to follow the match.
You don't need a big team or an Enterprise AWS budget for that. You need a well-designed event model, a lean serverless stack, open APIs, and genuine co-design with the people you're building for.
All of that fits inside the Free Tier. All of that fits in a weekend.
If you're interested in exploring this idea further, I have the full architecture design, cost tables, and validation roadmap. Happy to share.
And if you can enter the challenge — do it. The question deserves a good answer.
Misael Calisaya — AWS Student Builder Leader Group UCASAL · Salta, Argentina
Enjoyed reading this content? Let the author know!
Your likes, comments, shares, and saves help creators reach more builders.
Loading recommendations
Loading article