The rubric
The competition publishes its rubric: five categories, 100 points, six levels from 0 to 5, scored by Council members. A mini app is a web page inside Nimiq Pay and a web page cannot be a Bluetooth node, so NIMmesh cannot enter.
Claude Code produced the numbers and the reasons. The three columns and the decision to publish were mine. Nobody at Nimiq has scored this app. A row is worth 5 points unless it says otherwise.
Reading the score
| Harsh | Scored the way an unimpressed judge would, today. Every doubt resolved against the app. |
| Fair | The most defensible reading of the evidence that exists today. |
| After | After the funded three months: the mesh proven across Android and iPhone, swaps on mainnet, Bitchat on Android, Google Play. |
Score by category
- Harsh
- Fair
- After
One category, the promotion checklist, scores a post about a submission. That puts 5 of the 100 points out of reach, and the ceiling is 95.
The scorecard
| Criterion | Harsh | Fair | After |
|---|---|---|---|
Functionality, reliability and usefulness · 45 points | |||
Core featureWhy“Main app promise completes without error and does what it says.”Four mainnet transfers have crossed the mesh, three of them phone to phone with no computer in the path. All four were iPhones. Android has no receipt yet. | 3 | 4 | 5 |
Error handlingWhy“Failures are caught and explained clearly instead of crashing or freezing.”A method that is not built yet rejects by name rather than returning a plausible lie. A send that cannot broadcast fails loudly instead of pretending. | 3 | 4 | 4 |
SpeedWhy“Loads fast and responds immediately to input.”Native Rust core, no WASM and no consensus client. There is one public Nimiq RPC endpoint and it was down for a day. Not our bug, still the user's wait. | 3 | 4 | 4 |
StabilityWhy“Full flow runs end to end without dead ends, blank screens, or broken states.”The paid path runs end to end on iPhone. A send with no neighbour sits at pending with nothing to tap, which is correct and still reads as a dead end. Android has never completed the flow. | 2 | 3 | 4 |
CompletenessWhy“Feels like a finished product, not a prototype with missing pieces.”Sideload only, Android mesh unproven, swaps testnet only, LoRa and SMS still documents. It is an alpha, and every page says so. | 1 | 2 | 4 |
Real needWhy“Addresses an actual problem or want rather than a technical exercise.”If you have a phone but no data, you do not need your own connection. You need to stand near someone who has one. | 4 | 5 | 5 |
Target audienceWhy“Who this is built for is obvious within seconds.”The site says it in the first line. The app does not. Open it and you see a wallet with a grey status line, and nothing says the mesh is for people whose data is unreliable, censored or unaffordable. | 3 | 4 | 4 |
OriginalityWhy“A new idea or a meaningful improvement on something that already exists.”Not one of the 68 apps in Cycle I did offline relay. Every Bitcoin mesh bridge has to chunk a transaction and put it back together. A Nimiq one is 139 bytes and does not. | 5 | 5 | 5 |
Repeat valueWhy“Gives a reason to come back rather than open once.”It is your wallet, so yes. The mesh only earns its keep when you are offline next to someone who is not, and that is not every day for most people. | 3 | 4 | 4 |
| Functionality, reliability and usefulness subtotal | 27 | 35 | 39 |
Nimiq Pay and Nimiq integration · 25 points | |||
Payments are coreWhy“Wallet or payment function is central to the experience, not bolted on.”The product is a signed Nimiq transaction and nothing else. It is not Nimiq Pay, which is the name on the category, and the harsh column docks it a point for that. | 4 | 5 | 5 |
Payment statesWhy“Success, failure, and cancellation are all handled without leaving the user stuck.”Success comes back through the mesh as a receipt and the send flips from pending to settled. A failed broadcast is reported. There is no way to cancel a signed send that is waiting for a neighbour. | 2 | 3 | 4 |
Trustworthy flowWhy“User always knows what they are paying, to whom, and what happens next.”Amount, address and a real funds warning on the send sheet, in the wallet's own layout. What happens next is the weak part. Meshed and pending are the whole explanation. | 3 | 4 | 4 |
Mobile experienceWhy“Behaves natively inside Nimiq Pay with no zooming, cut-off buttons, or desktop layouts.”It is native, so nothing zooms or clips. It does not run inside Nimiq Pay and never will. The harsh column reads the sentence as written and gives nothing. | 0 | 3 | 3 |
NIM and ecosystemWhy“Uses NIM or the wider Nimiq ecosystem beyond simply accepting a payment.”Every byte on the mesh is a NIM transaction, the 139 byte size that makes it work is a Nimiq property, and the swap engine uses Nimiq's own hash time locked contracts. | 4 | 5 | 5 |
| Nimiq Pay and Nimiq integration subtotal | 13 | 20 | 21 |
Real usage · 15 points | |||
Unique usersWhy“25 or more users, 15 points. 11 to 24, 10 points. 4 to 10, 6 points. 0 to 3, nothing.”Nobody has counted, and the first band starts at four. Until there is a count this is zero in every honest column. The Google Play closed test in month two requires twelve testers for fourteen days, which is the 10 point band by Google's rule. | 0 | 0 | 10 |
| Real usage subtotal | 0 | 0 | 10 |
Design and UX · 10 points | |||
Visual qualityWhy“Clean, consistent, and professional enough that a stranger would trust it.”Nimiq's palette and type, used consistently across five languages. Almost none of it is invented. The grey mesh status line is the one piece a stranger would not recognise. | 3 | 4 | 4 |
Ease of useWhy“First-time user reaches the point of the app within 60 seconds without instructions.”Today: find the file, allow unknown sources, install, write down 24 words, then find a second phone. Nowhere near 60 seconds. Google Play deletes the first three steps. | 1 | 2 | 3 |
| Design and UX subtotal | 4 | 6 | 7 |
Builder promotion checklist · 5 points | |||
Skool community postWhy“Builder posted the Mini App in the Skool community at least once. 2 points.”There is no Mini App to post. Zero in every column, and not a judgement about the app. | 0 | 0 | 0 |
Social media postWhy“Builder shared at least one public social media post about their Mini App and the competition. 3 points.”Same. The post the checklist wants is about a submission, and there is none. | 0 | 0 | 0 |
| Builder promotion checklist subtotal | 0 | 0 | 0 |
| Total out of 100 | 44 | 61 | 77 |
The ceiling is 95
Against 95 the three columns read 46%, 64% and 81%.
Two more rows lean the same way: real usage is a user count nobody has, and mobile experience says inside Nimiq Pay. The rest of the gap is the work in the funding request.
The rubric changed
This page first went up on 29 August 2026 against a 105 point, 21 criteria rubric. By 2 September the competition had replaced it with the 100 point version scored above.
| Rubric | Harsh | Fair | After | Ceiling |
|---|---|---|---|---|
| 105 points, 29 August | 59 | 76 | 86 | 95, unique users and submission quality |
| 100 points, 2 September | 44 | 61 | 77 | 95, the promotion checklist |
Real usage became a 15 point tier on user count, Nimiq Pay integration its own 25 points, design fell to 10. The demo video row is gone; a video is proof now, not points.
Sources
Criteria and levels: miniappscompetition.com/scoring, read 2 September 2026. Prize schedule: the payout page. The transfers behind the functionality rows are on the home page.