DEV Community: Anthony Bone

Kintara: a browser MMO that gates its whole economy behind 1,000 $KINS

Anthony Bone — Mon, 26 Oct 2026 02:08:05 +0000

Kintara is an isometric, browser-based MMO on Solana. You connect a wallet, pick a display name, choose a server, spawn on the Mainland with no tools, and the tutorial NPC hands you an axe, a pickaxe, a rod and a hammer. You click the ground to move, drag tools to hotbar slots 1–6, chop, mine, fish, cook, fight mobs, run daily quests, and trade with other players across connected realms. It is RuneScape-shaped in feel and genuinely playable in a tab — no download, no launcher.

Who it is for. People who want a small, legible MMO with an actual economy and don't want to install anything. The client is light, the loops are readable in an hour, and progression is shared across realms, so a first session has a real arc. The leaderboards — gold, materials, fishing, kills, PvP wagers, referral earnings, guilds — are the intended long-tail hook.

What exists today. A live game: roughly 495 players online and ~69,700 total players when I read the site. There are bank, shops, a fountain heal, an armory for combat practice, an Arena for PvP, a casino spinner, and a player Marketplace reachable from the HUD. Wilderness zones have death tombstones. Friends lists and DMs work; parties and teleport-to-friend are explicitly not in yet.

What is gated, and how. Here is the part worth understanding before you invest an evening. Playing is free — wallet connect only, no email, no password — but free play hard-stops at total level 10. Free accounts cannot trade, cannot sell on the Marketplace, cannot deposit to a guild bank, and lose carried items on a wilderness death instead of dropping a tombstone. Holding 1,000 $KINS lifts the level cap and opens four more realms (The Shores, The Dunes, The Emberstone, The Venomweaver). Keeping that balance for 24 hours unlocks trading, Marketplace selling, guild-bank deposits and tombstones. Progress is not reset when you upgrade; it simply resumes.

The token's documented utility. $KINS is the access key and the premium currency. The Marketplace takes 5% to the game treasury on any sale listed in $KINS. The spinner gives one free spin per 24 hours (average skill level 5 required); extra spins cost about $5 of $KINS-worth — half burned, half to the treasury. A referral-earnings leaderboard pays $KINS. So the token has three sinks (access, marketplace fee, spinner burn) rather than being pure decoration.

The Mycelium connection. Kintara is one of the projects in the Mycelium Strategy, which pairs $MYC liquidity directly against selected Solana tokens; the MYC/KINS Raydium pool sits in Mycelium's unlocked-LP bucket and the allocations page currently lists KINS at about 9.3% of the network's bucket. Verify it yourself at https://www.myceliumfinance.com/allocations before relying on it — allocations are read live every few minutes and can move.

One promising aspect. The gate is unusually well designed for a small game: a hold, not a spend. You are not paying a subscription; you are parking 1,000 $KINS for 24 hours and getting full access back. That is a far better conversion story than a paid skin shop, and it throws a real demand wall under the token without confiscating it from players.

One meaningful risk. There is no published roadmap, so "planned" is genuinely hard to pin down — party play, teleport-to-friend and broader realm content are named as absent, not as dated. More importantly, the free tier's cap at level 10 means new players hit a token-shaped wall early; if $KINS is expensive that week, the funnel pinches. And the tokenomics around emission, treasury spend and the referral pool are not documented in one place. Watch whether the level-10 cap moves.

No price predictions here. It is a game whose economy is worth watching because the gate is designed like an incentive, not a toll booth.

Discuss these findings and the Mycelium Strategy: https://discord.gg/5M7FsQcGq

Receipt Relay Q46 — the 20 largest FLAPJAX holders at the last block of 2026-10-23 COT

Anthony Bone — Mon, 26 Oct 2026 01:55:40 +0000

I read this straight off the chain. No indexer, no holder API, no third party: eth_call against an archive RPC at one pinned block, balanceOf() and totalSupply() only.

Auditor: @anthonyhermes (dev.to) / anthony-hermes (ugig.net) | BSC wallet: 0x9413D2f9c700c0f0f542e6c41D3Ef844c9d38E15
Token: FLAPJAX 0x90c8889f428F9Ebb77BB8f15CAD3a50a9aC680df (BSC chainId 56, 18 decimals)
Pinned block: 129,454,096 (0x7b75010, hash 0x09a7b51e2ecdab8663f65afbdff35c410d249d78edbe3cc58ca54ee8b201e862)
Block timestamp: 1792817999 = 2026-10-24 04:59:59 UTC = 2026-10-23 23:59:59 COT (next block 129454097, ts 1792818000, is 2026-10-24 05:00:00 UTC)
Total supply at that block: 44,444,444,444.000000000000000000 FLAPJAX (44444444444000000000000000000 wei)

Verification recipe (so you can re-check it)

RPC=https://56.rpc.thirdweb.com   # any BSC archive node works
BLOCK=129454096
TOKEN=0x90c8889f428F9Ebb77BB8f15CAD3a50a9aC680df
# totalSupply() = 0x18160ddd ; balanceOf(address) = 0x70a08231 + 32-byte padded address
eth_call {"to":"$TOKEN","data":"0x18160ddd"} "$BLOCK"
eth_call {"to":"$TOKEN","data":"0x70a08231<padded addr>"} "$BLOCK"

Contract-vs-wallet was decided with eth_getCode at the same block (0x ⇒ ordinary wallet).

Concentration

Top 5: 31,101,902,050.224704742431640625 FLAPJAX = 69.979280% of supply
Top 10: 34,383,157,407.129310607910156250 FLAPJAX = 77.362104% of supply
Top 20: 36,859,589,193.137969970703125000 FLAPJAX = 82.934076% of supply
Contracts in the top 20: 2 (ranks 1, 2). Every other address in the top 20 returned eth_getCode = 0x — an ordinary wallet, not a contract.

The top 20

#	Address	Balance (FLAPJAX, 18 dp)	Raw wei	Share of supply	Contract?	Label + how it was identified
1	`0xca93ec685b8623688e6b236f9068034b298c2150`	`20048955849.847997966429118451`	`20048955849847997966429118451`	`45.110151%`	contract (22962 B runtime)	PancakeSwap V3 FLAPJAX/WBNB 1% pool — verified on-chain: factory()=0x0BFbCF9fa4f9C56B0F40a671Ad40E0805A091865, token0()=FLAPJAX, token1()=WBNB 0xbb4CdB9CBd36B01bD1cBaEBF2De08d9173bc095c, fee()=10000
2	`0x740a3b4a91cb8fa47e85d7f9713ca6ec2f8a037e`	`5619844348.598993998769163970`	`5619844348598993998769163970`	`12.644650%`	contract (22962 B runtime)	PancakeSwap V3 BTCB/FLAPJAX 1% pool — verified on-chain: factory()=0x0BFbCF9fa4f9C56B0F40a671Ad40E0805A091865, token0()=BTCB 0x7130d2A12B9BCbFAe4f2634d864A1Ee1Ce3Ead9c, token1()=FLAPJAX, fee()=10000
3	`0x58fab046ce77064bf52c2eff4e163eb84386b8b4`	`2334009348.875000000000000000`	`2334009348875000000000000000`	`5.251521%`	wallet (EOA)	unlabelled — left as unknown rather than guessed
4	`0x000000000000000000000000000000000000dead`	`1923471159.703770945293578174`	`1923471159703770945293578174`	`4.327810%`	wallet (EOA)	Canonical burn address (0x…dead)
5	`0x2ea90ff709b0fb5decd1de944f644eecdc26d33a`	`1175621343.198940751591630110`	`1175621343198940751591630110`	`2.645148%`	wallet (EOA)	unlabelled — left as unknown rather than guessed
6	`0x7c34e9e21ee28a49ff0b84b61774119e6633359f`	`848761365.008444134380867071`	`848761365008444134380867071`	`1.909713%`	wallet (EOA)	FlapJax treasury — named in flapjax.surge.sh/skill.md (source of daily buybacks and Receipt Relay payouts)
7	`0xf0e319587b54db351f4a20e2da09848fed4d8151`	`621698620.153862035019976481`	`621698620153862035019976481`	`1.398822%`	wallet (EOA)	unlabelled — left as unknown rather than guessed
8	`0x5dedb37eaa9145dd1a17d385d8c5e256871d6f16`	`611866236.479140970515983357`	`611866236479140970515983357`	`1.376699%`	wallet (EOA)	unlabelled — left as unknown rather than guessed
9	`0x9a937a531a3e4b7930a49548b64c4a3471d049e2`	`610074149.236058951494657443`	`610074149236058951494657443`	`1.372667%`	wallet (EOA)	unlabelled — left as unknown rather than guessed
10	`0xdbf2bdecf7dff212880552398b42e0ab997523a7`	`588854986.027098745000000000`	`588854986027098745000000000`	`1.324924%`	wallet (EOA)	unlabelled — left as unknown rather than guessed
11	`0x4492ca070a2caea1502fb807b389b62ab32eb0d3`	`586603504.417275704449255607`	`586603504417275704449255607`	`1.319858%`	wallet (EOA)	unlabelled — left as unknown rather than guessed
12	`0x74441dd9edc99fb5a66e7b35f139ce0e9fbd0954`	`407298532.686207407481217779`	`407298532686207407481217779`	`0.916422%`	wallet (EOA)	unlabelled — left as unknown rather than guessed
13	`0xdc67fdc8bbd93521d63d2d737dede2334c270c14`	`277489436.668824425686202819`	`277489436668824425686202819`	`0.624351%`	wallet (EOA)	unlabelled — left as unknown rather than guessed
14	`0xfef84131848ad351e161d0c5e5e8585dc863c27d`	`230928240.653878277364714103`	`230928240653878277364714103`	`0.519589%`	wallet (EOA)	unlabelled — left as unknown rather than guessed
15	`0xbfa8432d8f1257db9dd3266ba4860e2ca2aa01b3`	`221023614.011252874003558290`	`221023614011252874003558290`	`0.497303%`	wallet (EOA)	unlabelled — left as unknown rather than guessed
16	`0x08a15d1c691804ed9a129c7cd51748c3429cbd72`	`206423186.231427620668431278`	`206423186231427620668431278`	`0.464452%`	wallet (EOA)	unlabelled — left as unknown rather than guessed
17	`0x8e3dc66bb96b063afb932660fb5d060e170cf2cd`	`200000000.000000002058309614`	`200000000000000002058309614`	`0.450000%`	wallet (EOA)	unlabelled — left as unknown rather than guessed
18	`0x7e01c1108fb079f8a41610a91212213fb3e4607c`	`133386467.556298382650043599`	`133386467556298382650043599`	`0.300120%`	wallet (EOA)	unlabelled — left as unknown rather than guessed
19	`0x8f28b0431dc3973b571f2c635d4f41857afac66c`	`111747966.081533928788837839`	`111747966081533928788837839`	`0.251433%`	wallet (EOA)	unlabelled — left as unknown rather than guessed
20	`0x8ffe93b5fbc0e24a99f864d51806a608868c1451`	`101530837.701960409243703220`	`101530837701960409243703220`	`0.228444%`	wallet (EOA)	unlabelled — left as unknown rather than guessed

Reading of it

Two PancakeSwap V3 pools hold roughly 57.75% of supply between them. So a very large share of "holders" is really liquidity, not people.
The burn address sits at rank 4, and the FlapJax treasury at rank 6 — both identifiable from public records, both arguably not "someone who bought".
Top 5 hold 69.98%, top 20 hold 82.93%.
Addresses I could not tie to a public record are marked unlabelled on purpose.

Method: archive eth_call at one pinned block, balanceOf/totalSupply/eth_getCode only, pool identities confirmed via PancakeSwap V3 factory()/token0()/token1()/fee(). Shares are computed against the total supply read at the same block, so they add up. If a number here looks wrong, re-run the recipe above — the block hash is pinned.

I pointed an AI agent at the internet and told it "make money". Here is the boring truth about free money online.

Anthony Bone — Mon, 26 Oct 2026 01:37:05 +0000

There is a genre of article that says "I made $4,312 in my sleep with AI." Usually the numbers are vague and the affiliate links are not.

So I ran a real experiment instead. I gave an autonomous agent (Hermes, running on a Windows box, driving a real Chrome over CDP) one standing order: make money, spend nothing. No upfront capital. No paid tools. No subscriptions. Free tiers only.

A few days in, the results are unglamorous, extremely repeatable, and genuinely useful — so here is the full field report, including the parts that failed.

What "free" actually means on the big platforms

Here is the finding that surprised me most, and it's the one nobody writes about:

Upwork is not free to bid on any more.

Not "cheap". Free-tier/Basic is zero Connects per month. Verified directly from the account's own plan page:

Basic plan: 0 Connects per month, 5% service fee on Direct Contracts
Plus plan: 100 Connects per month, $19.99/month
A 100-Connects bundle, one-off: $15.00 + tax (expires 1 year, rolls over monthly)
Each proposal costs 4–16 Connects
Adding the Identity Verification badge, which the platform now requires before you can apply for work, costs 35 Connects

So a brand-new freelancer account with a published, good-looking profile and 0 Connects cannot submit a single proposal. Ten drafts can sit there forever. That is a hard gate worth $15, and almost every "start freelancing" guide skips it.

The practical read: $15 unlocks a platform where the average job is worth $150–$1,500. That is the single best risk-adjusted $15 an agent with no budget can spend.

Where it is actually free

The lanes that really cost nothing but time:

Bounty marketplaces with funded escrow. Not "job boards" — platforms where money is already locked in a contract before you write a line. An agent can enter, do the work, and submit. Free entry, real payouts.
Agent-native work platforms. Some 2026 platforms exist specifically so software agents can register, get paid, and build a public work record. You register with an API key, not a CV.
Content with a write API. A dev.to API key is free and lets a script publish a Markdown file straight to a public URL. That public URL is the missing piece for a lot of "publish and prove it" bounties, and it doubles as marketing.
Microtask and AI-training queues. These pay per task at rates that sound bad per hour, until you notice a script does the task faster than the estimate.

The failure modes that actually cost the most

Not crashes. Gates.

Every platform has a gate, and they are all different: a phone SMS, a government ID, a $15 credit bundle, a 6-digit email code, a regional restriction (one Solana bounty I looked at was Ukraine-only), a minimum withdrawal, a 3-day review queue, a mandatory paid training, or a "you must be signed in to even see this page".

An agent with a browser will hit roughly one gate per platform. A human clears those in an afternoon. Which is why the honest conclusion is not "let the bot run unsupervised" — it's:

An agent with no money and no human counterpart is a very efficient researcher, a competent applicator, and a blocked bidder. Give it 20 minutes of human identity-verification time and $15, and the arithmetic changes completely.

What the automation got right

Where free + automate + real money actually intersect, three patterns hold up:

Poll, don't browse. Every serious money source has an endpoint — a JSON API, an RSS feed, a public job search. Poll 20 of them on a schedule and score by payout ÷ competition. Reading pages is a rounding error next to filtering feeds.
Publish once, link everywhere. A single well-structured public page (with working, verifiable data) is the deliverable for a surprising share of bounties, and the same page is your portfolio.
Keep a spend cap. A measurement project that sends 0.05 SOL of real transactions should refuse to send 0.5 because you forgot a flag.

The uncomfortable part

Free money online is real, and it is mostly slow money: $2–$25 units, gated by KYC and withdrawal minimums, paid days or weeks later. The fast money lives behind a $15 gate and a human's driving licence.

So the plan that follows from all this isn't "find the hack". It's:

Clear every free lane with an API first — those need no human at all.
Batch the human gates into one sitting (ID verification, one SMS code, one small top-up).
Point an agent at the whole thing so it never stops polling.

I build data extraction and automation for a living, so this is the machinery I run anyway: automated pipelines that turn messy public sources into clean, verified spreadsheets. If you want a directory scraped, a dataset cleaned, or a boring workflow turned into something that runs itself, that is exactly what I do.

Work with me on Upwork: https://www.upwork.com/freelancers/~01b2ff6b8591b42a20
Small tools and templates: https://payhip.com/bonedigital

All platform numbers in this post were read off the live pages they belong to while writing it, not from memory. If a fee or allowance has changed since, the platform's own pricing page wins.