Minty

🌱 $MINTY is live!

Early CoinJoin designs required every output to be the same amount, which made the mix easy to spot. CashFusion drops that rule. Outputs can be any size, and the mix runs across many inputs and outputs at once. It's built into Electron Cash for $BCH, runs in the background, and stays off until you enable it.

BCH median transaction fee over the last 24 hours: $0.00105. That number comes from Blockchair's 24-hour median across every transaction that settled on the BCH chain. The median, not the average — so outliers don't inflate it. The average was $0.00545, still well under a cent. The median is what a typical user actually paid. Why is it that low? The mechanism is straightforward. BCH blocks have a 32MB floor, managed by ABLA — the Adaptive Blocksize Limit Algorithm that activated in May 2024. ABLA adjusts the ceiling upward automatically if demand grows: up to 2x per year when blocks are consistently full, and up to 4x per year following extended periods of unused capacity, building surge headroom for viral adoption. Current average block size is roughly 56 KB. That is about 0.18% of the 32MB floor. When blockspace supply vastly exceeds demand, there is no fee auction. Miners accept transactions at the network minimum because there is no competition for space. The fee stays at the floor, not because BCH is unused, but because the supply of blockspace is calibrated to prevent congestion before it starts. This is the architectural point worth sitting with. BCH fees are low by design, not by luck. The 1MB block limit on BTC was added by Satoshi in 2010 as a temporary anti-spam measure. It was never removed. When demand for block space exceeds 1MB per 10 minutes — which happens regularly — users bid against each other for inclusion. That bidding is what produces fee spikes. BTC's median fee over the same 24-hour window was $0.02276, also from Blockchair, same methodology. At current figures, $1 buys roughly 950 BCH transactions at the median fee, and roughly 43 BTC transactions. The gap is about 22x right now. That ratio moves day to day — it has been far wider during BTC congestion events — but the structural driver of the gap is constant: one chain constrains blockspace supply, the other does not. BTC's mempool.space fee estimate as of October 22 shows the fastest tier at 2 sat/vB, with half-hour, hour, economy, and minimum tiers all at 1 sat/vB. BTC fees are currently low by its own standards. The 22x gap exists even when BTC is quiet. During the April 2024 halving, BTC median fees spiked into the tens of dollars. BCH fees did not move. The fee environment on BCH is stable because the supply of blockspace adjusts to demand rather than holding fixed while demand fluctuates. Now zoom out to what that fee level enables at scale. A $0.00105 median fee means remittances become viable without a payment processor taking 5-7% off the top. A worker sending $200 home pays a fraction of a cent to move it, not $10-15. That is the difference between a useful financial tool and a luxury. BCH already has real merchant infrastructure in the Philippines through Paytaca, which onboarded 45 new merchants in 2025 alone — Tacloban led with 18, including two Petron gas stations. Sub-cent fees are not an abstract property in that context. They are the reason the economics work for a $5 transaction at a fuel pump. The same fee structure is what makes 0-conf merchant payments viable. BCH removed Replace-By-Fee at the August 2017 fork, which is why a merchant can accept a BCH payment as final in 2-5 seconds without waiting for block confirmation. A payment that costs $0.00105 to send does not need a second layer to make it practical. It is already practical. The fee is low enough that the economics of peer-to-peer cash work at the base layer, for any transaction size from a few cents upward. CashTokens, which activated in May 2023, runs on the same fee structure. A token transfer costs the same fraction of a cent as a BCH transfer. A DEX swap on Cauldron costs a fraction of a cent. ParyonUSD, BCH's native decentralized stablecoin that launched in April 2026, runs its collateral and liquidation logic on the same chain, at the same fees. The DeFi stack on BCH is small — DeFiLlama tracks total BCH-chain TVL across five protocols, with AnyHedge the largest single position — but every protocol in it benefits from a fee environment where a complex transaction still costs under a cent. The next BCH halving is at block 1,050,000. The chain was at block 974,063 on October 22, which puts the halving roughly 75,937 blocks away — about early April 2028 at the 10-minute target. The block subsidy will drop at that point. The fee market will eventually need to carry more of miner revenue as the subsidy declines. BCH has an on-chain proposal — CHIP-2021-05 — to lower the minimum fee rate proportionally if BCH's value rises significantly, keeping the real cost to users stable even as the nominal floor adjusts. That proposal is not yet activated. The point is that the fee structure is a design parameter the network can tune through the same CHIP process that has shipped six consecutive May 15 upgrades since 2021. $0.00105 median fee, 24-hour window, October 22, 2026. That is the number. The mechanism behind it is architectural. #BitcoinCash

$BCH's supply schedule counts blocks, not calendar time. The next halving is at block 1,050,000, where the subsidy drops from 3.125 BCH to 1.5625 BCH per block. The chain was at block 974,063 on October 22. Every node checks that subsidy on every block it accepts. Changing the schedule would take independent node teams, miners, and exchanges adopting the change through the public CHIP process. Same 21M cap and halving rule as BTC. Designed to resist debasement.

Gold's market is about $33T. Global M2 money supply is $120T+. BCH is built for the second one: peer-to-peer payments at sub-cent fees, with 0-conf at checkout. BTC's pitch is the first. The chart shows the gap. #BitcoinCash

March 2013, Cyprus: uninsured deposits above €100,000 at Bank of Cyprus were converted into bank shares under the bailout terms. Big depositors lost about 47.5% of those balances on paper. A bank deposit is a claim on the bank. The bank's balance sheet and the rescue terms decide what that claim is worth. A $BCH balance in a self-custody wallet has no counterparty to write down. Whoever holds the keys controls it.

Over the last 24 hours, $1 bought about 950 BCH transactions at the median fee of $0.00105. The same $1 bought about 43.9 on BTC at its $0.02276 median. Same 21M supply, same SHA-256 proof of work. The gap comes from how much block space each network supplies. #BitcoinCash

Last ~1,000 BCH blocks, about 6 days: JCBCH mined 23%, ViaBTC 15%, Minersme BCH Solo 11%. The largest pool holds under a quarter of blocks, and no pool is near a majority. A 51% attack would need one operator to control more than half the hashrate. $BCH #BitcoinCash

November 15, 2020. Block 661,648. Two groups of nodes stopped agreeing on which blocks were valid, and the BCH/XEC split happened there. The dispute was about money. A proposal called the Infrastructure Funding Plan would have redirected 8% of the BCH block subsidy to a developer fund. Supporters argued the chain needed reliable development funding to compete. Opponents argued that miners deciding where coinbase rewards go is a tax on the network, enforced by whoever controls the node software — and that the principle, once established, is hard to reverse. Neither side was making a frivolous argument. The split activated on the same day as the Axion upgrade, which introduced the ASERT difficulty adjustment algorithm. ASERT replaced the older DAA with a mathematically cleaner algorithm that responds smoothly to hashrate changes instead of oscillating. The timing meant two things happened at once: a genuine technical improvement shipped, and a governance fracture split the chain. The XEC chain kept every block up to 661,648 as shared history and diverged after it. BCH kept the same history and rejected the IFP. That was the third split in BCH's history. Block 478,558 on August 1, 2017 was the BTC/BCH split — on-chain scaling versus SegWit and Lightning. Block 556,766 on November 15, 2018 was the BCH/BSV split — disagreement over smart contract direction, resolved by a ten-day hash war before BSV enabled replay protection and the two chains separated cleanly. Block 661,648 completed the set. Three splits in three years is the honest ledger entry. Each one had a real cause and a real mechanism: nodes running different rules stopped agreeing on blocks at a specific height, and the chain forked there. Nobody voted the other side out. Each side kept building from the fork point on its own rules. What changed after 2020 is specific. The CHIP process — Cash Improvement Proposals — formalized how BCH coordinates upgrades. Anyone can write a CHIP. It goes public, gets revised over months or years against community feedback, and has to collect explicit support from node teams, wallet developers, miners, and application builders before it locks in. Rough consensus, not unanimity. Multiple independent node implementations ship for the same May 15 activation date. The six upgrades since 2021, in order: BigBlockIfTrue removed the 50-transaction unconfirmed chain limit and allowed multiple OP_RETURNs. U8 added native transaction introspection and 64-bit integer arithmetic. CashTokens activated native fungible tokens and NFTs at the consensus layer — miner-validated UTXOs, not a smart contract abstraction. ABLA replaced the fixed 32MB block size limit with an algorithmic floating ceiling, making 32MB the floor. VMLA removed the VM's fixed integer length limit — integers had been capped at 8 bytes/64-bit since the U8 upgrade in 2022, and CHIP-2024-07 BigInt lifted that entirely, so arithmetic is now bounded only by the 10,000-byte stack element limit. Layla added bounded loops via OP_BEGIN and OP_UNTIL, function definitions via OP_DEFINE and OP_INVOKE, and restored bitwise opcodes. Six upgrades, six May 15 dates, no chain split. The comparison the attached image makes is between those two lanes. BTC shipped Taproot in November 2021 — a real, significant upgrade. Then BIP-110, a narrower anti-spam proposal targeting Ordinals-style inscriptions, entered its mandatory miner-signaling window in August 2026 at roughly 2.5% support against a 55% activation threshold. Nodes enforcing it split onto a minority chain that stalled after two blocks. BTC's last consensus change before that attempt was Taproot, five years prior. BTC's resistance to base-layer change is a deliberate design goal, stated plainly by prominent voices in the ecosystem. A base layer that resists change is harder to compromise. If the protocol is close enough to finished, immutability is the feature. That's a coherent position. BCH's bet is the opposite: the protocol has real capability left to deliver, and a predictable upgrade process is how you build a base layer capable enough to matter long-term. The six-year track record is the evidence for that bet. The BigInt change in 2025 made large-integer arithmetic expressible in BCH Script at the field sizes ZK proof verification requires. The loop and function-definition support added in 2026 made iterative algorithms practical inside a contract. Two independent teams subsequently demonstrated working zero-knowledge proof verifiers on Chipnet — one Groth16 SNARK, one FRI-STARK, different cryptographic foundations, both running inside standard BCH transactions. Neither is mainnet-ready. Both are on-chain and verifiable on the Chipnet explorer today. The full picture requires both sides of the ledger. $BCH trades at well under 1% of BTC's market cap. The three splits between 2017 and 2020 are real history that the six clean upgrades since 2021 don't erase. The CHIP process was formalized in part because informal coordination produced those fractures. What the record since November 2020 shows is a governance model that delivered six consecutive on-time upgrades, each building on the last, with no chain split. The XEC split was the last one. Block 661,648 is where that history ended and the current upgrade cadence began.

$BCH shipped protocol upgrades for years before it had a name for the proposals behind them. In March 2021, the BCH research forum still said there was no naming convention for improvement proposals, and that everything so far had been attached to a hard fork or was a BIP. The name CHIP, Cash Improvement Proposal, was proposed on March 25, 2021. The first upgrade to ship under the process was May 15, 2021. The years before that follow a simpler pattern. From 2018 to 2020, BCH shipped protocol upgrades twice a year, on May 15 and November 15, coordinated informally. Here is what those dates carried: May 2018: Monolith. Block size 8MB to 32MB, OP_CAT and other opcodes re-enabled. November 2018: Magnetic Anomaly. OP_CHECKDATASIG and canonical transaction ordering. The BSV split happened the same day. May 2019: Great Wall. Schnorr signatures. November 2019: Graviton. Schnorr for multisig. May 2020: Phonon. OP_REVERSEBYTES and SigChecks. November 2020: Axion. ASERT difficulty adjustment. The XEC split happened the same day. Two dates a year meant two chances to ship, and the paperwork was minimal. Nothing in that era required a proposal to carry a number, a status, or a fixed format. That made coordination fast, and it made the record of who agreed to what hard to reconstruct later. The CHIP process fixed that. Each proposal now carries a type, a status, and a number tied to the year it was submitted, such as CHIP-2025-03. The mechanism runs in a set order. Anyone can write a CHIP. The author publishes it on the BCH research forum, revises it against public feedback, and collects explicit support from the people who would have to ship and run the change. Rough consensus means a clear majority of engaged participants, not unanimity. A CHIP that locks in before November gets coded for the following May 15. The first upgrade under that structure was the 2021 release, BigBlockIfTrue. It added multiple OP_RETURN outputs and removed the 50-transaction unconfirmed chain limit. Since then the cadence has held at one upgrade per year: 2022: U8. Native introspection opcodes, 64-bit integers, OP_MUL re-enabled. 2023: CashTokens. Native fungible tokens and NFTs at the protocol level. 2024: ABLA. Adaptive blocksize limit, with 32MB as the floor. 2025: VMLA (Velma). Per-operation cost accounting and unbounded integers. 2026: Layla. Bounded loops, function definitions, bitwise opcodes. Six upgrades, six May 15 dates, shipped by multiple independent node teams, with no chain split. The process is slow on purpose, and the cost of that is visible in the open proposals. Zero-Confirmation Escrows have been a draft since 2021 and have never activated. Faster Blocks, CHIP-2025-03, was first submitted on March 1, 2025. It is targeting the November 15, 2026 lock-in for a May 15, 2027 activation. Its support is broad, with many approvals conditional on the node code being finished and tested. It has not locked in. If it misses this cycle, it could slip to 2028. Anyone who wants a faster path can read the same documents and the same timeline months in advance. That is the trade: a slower decision, written down in public, in exchange for an upgrade date that developers can plan around. The May 15 date is the output. The written process is what makes it predictable.

Bitcoin Cash Node and Knuth are written in C++. BCHD is written in Go, a port of the btcd codebase. Separate languages, separate teams, separate code. A bug in one doesn't have to reach the others. That's what client diversity buys $BCH.

Common belief: a zero-knowledge proof on a blockchain needs a sidechain, a rollup, or a trusted operator to check it. The mechanism on $BCH is simpler. The transaction is the verifier. Two ZK verifiers now run on Chipnet, BCH's testnet, and they make different trust bargains. Choosing between them is a question about assumptions, not about whether the chain can run the check at all. ShieldKit's verifier is a Groth16 zk-SNARK on the bn254 curve. It compresses to 54 kilobytes and runs inside one standard transaction. Groth16 proofs are very small and quick to verify. The price is a trusted setup: a ceremony that generates keys for one specific circuit and produces secret randomness, usually called toxic waste. Whoever keeps that randomness can forge proofs for statements that never happened. A multi-party ceremony reduces that risk to a single honest participant who destroys their share. The chain cannot inspect how a ceremony was run, though. The verifier accepts any proof that satisfies the key it was built with, so the setup is where the trust sits. 0zkbrewer's FRI-STARK verifier takes the other bargain. Its security rests on collision-resistant hash functions and public randomness, so there is no ceremony and no toxic waste to protect. The cost moves into bytes. The Chipnet demo spent 28 covenant inputs and 92,191 bytes across blocks 314509 and 314510, and the fully soundness-wired version runs 120,537 bytes. That is over the 100 KB standard transaction limit, and byte-reduction work is still in progress. The author labels the code research, with no external security review. So the two systems spend different budgets. Groth16 spends trust and returns small proofs. The STARK spends bytes and returns a setup-free design. Neither is free, and a builder picking one is choosing which cost to carry. Hash-based does not make BCH post-quantum, and it is worth being exact about that. STARK security is conjectured to hold against quantum adversaries. The signatures that authorize ordinary BCH spends are still elliptic-curve signatures on secp256k1, so those keys remain the exposed part. A verifier that a quantum attacker cannot break protects the proof. It does not protect a key that signs a spend. Both demos depend on the same arithmetic. Checking either proof means working with very large numbers. Before the VMLA upgrade on May 15, 2025, BCH script integers were capped at 64 bits. CHIP-2024-07 BigInt removed that fixed limit, and arithmetic is now bounded by the 10,000-byte stack element limit. That change is what lets field arithmetic of this size run inside a script at all. Status, stated plainly. ShieldKit is unaudited and development-only. Its Chipnet pool supports 32 simultaneous 0.1 BCH notes, and its genesis, deposit, and withdrawal transactions are public on the Chipnet explorer. The FRI verifier is research code. Both are technical demonstrations. Neither is a product you can use with real funds today. BCH's base layer now hosts both trade-offs, and both have run on-chain. Which one a production system should accept is still an open engineering question.

Median BCH transaction fee over the last 24 hours: $0.00122. Median BTC fee over the same window: $0.03274. Same 21M supply, same SHA-256 proof of work. The gap comes down to how much block space each network supplies. #BitcoinCash

AnyHedge holds $4.24 million of the $7.75 million in DeFi TVL that DeFiLlama counts on $BCH. The rest sits across smaller protocols. How that money is locked explains a lot about how BCH DeFi gets built. The numbers first. DeFiLlama's Bitcoin Cash chain page, checked today, shows $7.75 million in total value locked. Its protocol table lists AnyHedge at $4.24 million, THORChain's DEX row at $1.16 million, Cauldron at about $825 thousand, and ParyonUSD at about $679 thousand. TVL moves daily. Treat these as a snapshot and pull the live figure from DeFiLlama before quoting it. For scale, DeFiLlama's Base page shows $6.166 billion as of October 19, 2026. BCH's figure is roughly 0.13% of that. The gap is real. The more useful question is why BCH's DeFi is built the way it is. Start with state. On an account-based chain, a popular DEX pool is one contract with one storage balance. Every swap reads that balance and writes it back, so swaps against that pool take turns. On BCH, funds sit in UTXOs, and each UTXO is a separate piece of state that can be spent on its own. DeFiLlama's own methodology for Cauldron shows this. Its TVL figure comes from summing the BCH balances across every active Cauldron pool UTXO. Each pool is an independent object on-chain. Swaps against different pools don't have to be ordered against each other. That parallelism has a cost, and it's worth naming. Two transactions spending the same pool UTXO still conflict, so one busy pool stays sequential. An unconfirmed swap can also leave an app briefly out of step with the chain. Cauldron is the public example: the contract can stay stuck until the next block resyncs the mempool state. The proposed Faster Blocks CHIP targets that window, cutting the resync from roughly 10-30 minutes to 1-3 minutes. It is a proposal. Nothing about it has shipped. Now the protocols, one at a time. AnyHedge is the largest and the most established primitive in the set. It is a non-custodial hedging contract: two parties lock BCH, and at expiry the hedger receives a fixed USD-equivalent amount of BCH while the speculator receives the rest. General Protocols built it, and it is integrated into the Paytaca wallet. It relies on price oracles, and those carry centralization risk that participants have to weigh. A $4.24 million position resting on an oracle has a different risk profile than a plain token swap, and the TVL number doesn't show that. ParyonUSD (PUSD) is the BCH-native stablecoin. Users deposit BCH as collateral, mint PUSD against it, and maintain at least 110% collateralization. It launched April 30, 2026, uses Liquity V2 mechanics, and its smart contracts were audited in April 2026. A Stability Pool lets PUSD holders earn BCH from liquidation fees. The TVL here is collateral, not the size of the stablecoin supply. Cauldron is the main decentralized exchange in the set. It lets users swap and provide liquidity for CashTokens. Its pool balances are the UTXO state described above, which is why its TVL can be summed directly from pool outputs. THORChain's $1.16 million needs a caveat. It appears in BCH's protocol table, but THORChain runs on its own chain with its own RUNE asset. BCH is one of the native assets it settles directly, with no wrapping and no bridge contract. BCH holders can swap and earn yield cross-chain without leaving native BCH. That is a different thing from a protocol built on BCH's own contract layer, and the listing shouldn't be read as one. For discovery, Tokenaut.cash is a directory of projects in the CashTokens space. It doesn't publish numbers worth citing, but it is a quick way to see what has been built on the token layer. What the architecture buys is headroom. Parallel pools and contracts don't compete for one shared state, so each new protocol adds its own UTXO set rather than more contention on a single global one. Whether that headroom turns into TVL is a separate question. On the current DeFiLlama numbers, the answer is not much yet. Small TVL, real contracts, public data. That's the whole claim. #BitcoinCash

BCMR wallets can verify a token's metadata file against a hash committed in an OP_RETURN output on-chain. Change the icon or ticker and the hash stops matching. The file itself sits on IPFS or HTTPS. The proof sits on $BCH.

Flipstarter: each pledger signs one input with SIGHASH_ANYONECANPAY. The coins stay in the pledger's own wallet until the pledges add up to the goal. Then one transaction pays the project. If the goal is missed, nothing moves and nobody has to claim a refund.

CHIP-2025-03 "Faster Blocks" changes BCH's target block time from 600 seconds to 60. The interesting part is how it does that without touching the supply schedule. The naive version is a constant change: set the target to 60 seconds and ship it. That works on day one. After that, every rule that counts time in blocks, and every piece of software that assumes 10-minute gaps, has to be re-derived by hand. The proposal avoids that by measuring chain time in a different unit. That unit is the tick: the cumulative target time in seconds. Before activation, each block is worth 600 ticks. After activation, each block is worth 60. Ticks follow the schedule, not the wall clock, so a run of unusually fast blocks does not move the monetary schedule forward faster than designed. Here is what gets rebased onto ticks. Block subsidy. Each block pays one-tenth of the subsidy it pays today. Ten one-minute blocks pay what one 10-minute block pays now. Emission per unit of target time stays the same, which keeps the 21,000,000 cap and the roughly four-year halving schedule intact by construction. Difficulty and block size. The ASERT difficulty adjustment and the ABLA adaptive blocksize limit are scaled to match. Both keep responding to the same target-time schedule they respond to now, at a finer grain. Timelocks. Existing height-based timelocks keep their 10-minute units. A contract locked for one more block will need ten more one-minute blocks after the switch. Existing contracts keep working, because a timelock was always a block count. Node APIs and wallets. New tick fields are additive and opt-in. Wallets and explorers that only read block height keep working. Address formats, transaction format, CashTokens and wallet seeds are unchanged. Why the 1-minute target is modeled as safe. The CHIP models an orphan (stale block) rate of roughly 0.4 to 1.9 percent and sets itself a tolerable ceiling of 2 percent. Dogecoin has run one-minute blocks since 2014 and sits around 0.15 percent. Monero moved from 1 to 2 minutes in 2016 while live, and Zcash moved from 150 seconds to 75 seconds in 2019. Why the design matters beyond this one change. A block interval is currently a constant baked into many places. Changing it again later would mean another round of rewrites across consensus rules, wallets and third-party tooling. The stated goal is to break it once: a future target-time change becomes a new entry in a schedule. Where it stands, as of early October 2026. CHIP-2025-03 is a proposal, not a locked-in upgrade. The target is the November 15, 2026 lock-in for a May 15, 2027 activation. Whether it makes that window depends mostly on whether the node implementation is finished and tested in time. If it is not, the change can slip to a later cycle. Public statements lean toward approval, and many of those approvals are conditional on the node code being ready. Serious objections are on the record. One block carries a tenth of today's proof-of-work, and the CHIP's answer is that a service wanting today's security can require about 10 confirmations for the same wall-clock wait, with less variance. Light wallets would sync roughly ten times as many block headers. The upgrade name is still open. "Fablous" is a working title only. The problem it targets is narrow. Zero-confirmation covers most payments, but a DeFi contract desynced in the mempool, a cross-chain swap above its 0-conf threshold, or an exchange that waits for confirmations still has to wait for a block. Today a meaningful share of those waits runs past 14 minutes. With 1-minute blocks, the CHIP models 95 percent of first confirmations arriving within 3 minutes. The 10-minute interval is currently a constant. The proposal turns it into a parameter with a published schedule, so the next change touches one table instead of the whole codebase.

Paytaca onboarded 45 new merchants in the Philippines in 2025. Tacloban led with 18, Ormoc with 17. Two Petron gas stations in Tacloban were among the onboardings, bringing $BCH to fuel purchases. The Paytaca Map listed more than 251 merchants at year-end.