Extended UTXO
State lives in discrete boxes, not a shared global account. A transaction declares exactly which boxes it consumes, making dependencies explicit and validation locally predictable.
Programmable money, built from first principles
Bitcoin-inspired security, expressive eUTXO contracts and privacy-preserving proofs — engineered for useful finance without trusted intermediaries.
Proof of Work · launched 1 July 2019 · no ICO · no VC allocation
The eUTXO model
Every unspent output is a box: it carries ERG, native tokens, a guarding script and typed data. Registers R0–R3 are mandatory; R4–R9 are yours.
Inputs are destroyed, outputs are made. Because a transaction names exactly which boxes it touches, validation stays local and predictable — no global account to reason about.
A data input is referenced without being spent. Any number of transactions can read the same oracle box in the same block, so nobody has to win a race to use shared state.
Sigma propositions
ErgoScript is built on Sigma protocols: non-interactive proofs of knowledge of a discrete logarithm. Spending proves a statement rather than revealing a secret.
Join them with AND, OR and threshold operators. Ring signatures and multi-signatures are not bolted on — they are what the language already says.
atLeast(2) is enforced by consensus, not by an application convention. Only the branches that actually satisfy the statement need to be revealed.
The foundation
Ergo starts with the battle-tested UTXO and Proof-of-Work ideas, then extends them deliberately — without turning validation into an open-ended computation.
State lives in discrete boxes, not a shared global account. A transaction declares exactly which boxes it consumes, making dependencies explicit and validation locally predictable.
A typed, deterministic language for spending conditions. Each script is bounded; long-running computation is expressed as auditable sequences of transactions — multi-stage contracts.
A memory-hard Proof-of-Work algorithm designed for commodity GPUs. Its original non-outsourceability goal was removed in v2 after contract-based workarounds made pool resistance impractical.
The eUTXO model
Every unspent output is an Ergo box. Mandatory registers R0–R3 hold value, guarding script, tokens and creation metadata; optional densely-packed R4–R9 hold typed application data.
Tokens are first-class box contents rather than balances maintained by a separate smart contract.
A transaction can read a box without consuming it. Many transactions can reference the same oracle state without fighting to spend it.
A contract validates from declared inputs, outputs and context. Independent boxes can be processed concurrently; a hot shared box still creates contention.
Users build the transaction off-chain; miners verify a bounded proof. Computation need not become permanent global execution.
Programmable authorization
ErgoScript composes Sigma propositions: discrete-logarithm knowledge proofs joined with AND, OR and threshold logic. This supports multisignatures, ring-style authorization and zero-knowledge protocols at the language level.
A script is not an unlimited virtual machine. Expressiveness emerges across transactions, while each validation remains finite and costed.
Non-interactive proofs authorize spending without exposing the witness. Threshold policies can require k-of-n participants.
A box encodes one protocol state; spending it creates the next. Chaining these state transitions can express general computation without unbounded work inside one block.
Compact proofs summarize accumulated Proof of Work for light clients and cross-chain protocols. Ergo nodes support NiPoPoW bootstrap; integrations vary by product.
Permissionless block production
Proof of Work asks for a machine and electricity. There is no registry to join, no stake to bond, no identity to reveal — a new miner simply starts hashing.
A winner broadcasts and the network verifies. Participation cannot be revoked, because there was never a list you were on.
When block production runs through a known set and its relays, every hop is somewhere a transaction can be delayed, reordered or dropped — and somewhere an operator can be identified and pressured.
Consensus
This is an argument about architecture, not tribalism. If the goal is financial infrastructure that keeps working when someone powerful would rather it did not, the question is simple: how many parties must cooperate for a payment to be included, and can any of them be identified and pressured?
A miner assembles a block and broadcasts it. There is no relay, no builder, no committee round between a valid transaction and a block that contains it. Every additional layer in that path is a place where censorship can be applied.
Stake-based consensus needs a known validator set — enumerable, addressable, and in practice concentrated. Proof of Work has no such set. A miner can appear anonymously, contribute, and vanish, without ever registering.
Fewer moving parts means fewer trust assumptions and less that can quietly break. A consensus rule you can hold in your head is one you can audit — and one an independent implementation can actually reproduce.
Stake systems tend to route rewards toward existing stake, so capital that already holds influence acquires more. Work must be re-spent every block; yesterday’s hashrate confers no authority today.
Ergo launched with (weak) non-outsourceability, likely the first Proof-of-Work coin to ship it, specifically to make mining pools hard to form. In practice it carried real drawbacks and pooling emerged anyway through contract-based workarounds, so Autolykos v2 removed it. It is history worth knowing rather than a feature to claim.
Lithos pursues the same goal by a different route: a blockchain-agnostic protocol using Stratum as its networking layer, evaluating Non-Interactive Share Proofs rather than ordinary pool shares, with collateral contracts that let lenders earn yield backing pools and miners insert transactions into blocks directly. Status: testnet. First public releases arrived in November 2025 and require a fully synced node and Java 11. Not production software.
Storage rent
A box left unspent for four years becomes eligible for storage rent, charged per byte by the storage-rent subprotocol.
Miners may take the rent — roughly 0.14 ERG per four years for a plain box — or spend the box outright if its value cannot cover it.
Abandoned state stops being a permanent cost to every node and starts paying for network security instead. On-chain garbage collection that funds itself.
Long-term design
Most chains treat storage as free forever: write once, and every future node carries it. Ergo prices it. A box untouched for four years becomes eligible for storage rent — an idea closer to demurrage than to a transaction fee.
The fee is per byte, set by the storage-rent subprotocol, so a box pays in proportion to the state it actually occupies. For a box with no tokens and no complex script that works out at roughly 0.14 ERG every four years.
If the box does not hold enough ERG to cover the rent, a miner may spend it outright. This is the part that makes it garbage collection rather than a levy — dead state does not linger, it is cleared.
If the ERG in a box cannot pay the rent, the miner can claim the assets inside it — including NFTs and stablecoins. This is a genuine risk to users who park valuables in a box and forget it. Keep enough ERG in boxes you care about.
Rent supplements transaction fees once scheduled block emission runs down, giving miners a revenue source that does not depend on new issuance. It is a mechanism, not a promise that future revenue will be sufficient.
A small holding cost on idle state nudges coins back into circulation. This is deliberate monetary design in the demurrage tradition, not an accident of the fee model.
Nautilus has box consolidation built in and warns you when your UTXO set needs it; TokenJay offers a consolidation tool that reports box age. Consolidating resets the clock and is the standard mitigation.
Four pillars
State-independent money needs a neutral ledger, stable value on top of it, a way in and out, and money people create for each other — each piece trust-minimized on its own.
No pillar depends on the others working. One failing does not bring down the rest — that is what trust-minimized actually buys.
kushti’s claim is that Ergo now has all four at once — the first ecosystem where that has been true.
kushti’s framing
Ergo founder Alexander Chepurnoy (kushti) describes state-independent money as needing four layers, each trust-minimized. His claim is that the Ergo ecosystem now has all four at once — the first time that has been true anywhere.
Programmable, trust-minimized, nobody’s liability. Ergo and ERG: Proof of Work, launched with no ICO, no VC allocation and no conventional premine.
Stable value and financial instruments built as contracts over that reserve, not as a promise from a company. SigmaUSD, Gluon, DexyGold and USE.
Getting in and out without a custodian standing in the middle. Rosen Bridge, plus peer-to-peer on- and off-ramp markets made viable by trust-minimized insurance.
Medium-of-exchange money issued by people to each other, with reserves used only to reduce trust rather than to grant permission. ChainCash and Basis.
Pillars two and four in particular contain systems at very different maturity levels — see each section for status.
Stable value
Every design here holds real ERG reserves rather than an uncollateralized algorithmic peg.
One side is insulated from the swings; the other absorbs the volatility in exchange for the upside.
Push the reserve too thin and the volatile side hits its floor first — the stable side is protected by construction, not by promise.
Stable value
None of these are uncollateralized algorithmic pegs of the kind that failed elsewhere. Each holds real reserves and differs in how it prices, splits and defends them — DexyGold and USE in particular are two pegs built on one shared protocol, not two competing designs.
The first UTXO-based stablecoin, an instantiation of the AgeUSD protocol whose economic model was designed with IOHK, Ergo and Emurgo. Conservative reserve settings mean it avoids liquidations entirely: SigUSD holders are insulated while SigRSV holders absorb volatility in exchange for upside. Overcollateralized, and running through market crashes since early 2021.
Designed by Bruno Woltzenlogel Paleo with the Djed Alliance. Rather than issuing new tokens it splits an existing one into stable and unstable components, borrowing the metaphor from nuclear physics: fission divides ERG into stablecoin and reservecoin, fusion recombines them, and beta decay converts between them like an AMM swap. Conceptually the dual of a liquidity pool. The live product is gold-pegged: GAU tracks one gram of gold, GAUC tokenizes the reserve surplus with leveraged volatility and yield. Both fully backed by ERG.
The first implementation of the Dexy protocol: an algorithmic bank paired with an on-chain AMM, oracle-priced so one DexyGold tracks one troy ounce of gold, collateralized entirely in ERG. Public mainnet beta since July 2025 — the same framework USE below builds its USD peg on.
The Dexy protocol’s USD-pegged implementation — DexyGold’s algorithmic-bank-and-AMM design, applied to the dollar instead of gold, on a conservative 1:1 reserve model. The reasoning is pragmatic: centralized dollar stablecoins anchor most on-chain liquidity and carry censorship risk, but refusing to interoperate isolates Ergo and costs ordinary users. Built for cost efficiency — bridge fees targeted near 0.1%, a dedicated AMM instance at 0.05% within a tight band, six decimals to match major venues. StableMiner mints it locally from a node wallet, and published x402 payment examples use Babel fees, so a payer can spend USE without holding ERG for the miner fee.
Bruno Woltzenlogel Paleo — who designed Gluon partly in response — has catalogued the weaknesses of the AgeUSD family: sensitivity to the oracle, reserve drainage by large holders, reserve-coin holders being unable to sell below the collateral threshold, and the zero-equity problem. Ergo’s stablecoins are engineering, not magic, and they are discussed publicly as such.
Mutual credit
In ChainCash, anyone can issue an IOU note. It is money because someone accepted it, not because an issuer was authorised.
As a note travels it accumulates signatures. Its backing is the collective trust of everyone who passed it on — the more it circulates, the more it is worth trusting.
On-chain reserves can be attached to reduce trust where trust runs out. You start with people and reach for the blockchain only when you need it.
ChainCash & Basis
ChainCash is a monetary system for elastic money creation, combining human trust with optional blockchain-backed reserves. Its ambition is unusual: let people create credit among themselves without knowing anything about Ergo, and reach for on-chain assets only where trust needs reinforcing.
Anyone can issue an IOU note. When a holder spends it, they co-sign — so a note that has circulated widely carries the backing of everyone who accepted it. Its worth reflects a web of trust rather than a single issuer’s balance sheet.
Fixed-supply assets are poor media of exchange precisely because supply cannot answer demand. ChainCash lets money expand and contract with real economic activity, without a central bank deciding when.
Basis holds mutual-debt state with minimally-trusted trackers, committing it on-chain as an AVL-tree digest. Ergo contracts handle redemption and prevent double-redemption, so ordinary small payments never need to touch the chain. Debt is transferable with the issuer’s consent.
On-chain reserves are optional collateral a participant can attach to make their notes more acceptable. If a tracker goes offline, the reserve owner can withdraw after a configured delay.
The spender-signed model was described in Kenji Saito’s Peer-to-Peer Money: Free Currency over the Internet, and adapted to the Ergo setting. Work on top includes Agent Credit and experiments in agent-to-agent payments.
Be clear-eyed about this one. ChainCash and Basis are active research and prototype software, not a finished wallet-native payment rail. Contracts and APIs changed frequently through 2026, and some transfer paths still require raw Schnorr signatures, so normal wallet support remains constrained. A Basis whitepaper went out for review in June and a RAMICS-26 paper and deck exist. Treat published details as a moving reference and verify against the repositories before building on it.
kushti’s vision for the ecosystem
Ergo founder Alexander Chepurnoy has set out how he sees the ecosystem developing. It is worth reading as a stated direction from one person rather than a roadmap the protocol guarantees — which is roughly how he frames it himself.
His read on the wider market is bleak and specific: hacks on roughly a daily basis, token launches dead, most venture capital dead, vested insider allocations that retail will not absorb, and conferences with only one topic — institutionalization, which in practice often means replacing real Bitcoin with a paper claim on it while tokenizing traditional products for exposure. Meanwhile the original demand is growing: privacy concerns are rising, fiat currencies keep failing around the world, and financial markets in many places look fragile.
Proof of Work is already a peer-to-peer market: anyone may generate a block, with no registry to join and no operator to pressure (see Proof of Work above). Lithos extends that market to mining itself rather than replacing it.
Existing p2p ramp markets for ERG, Bitcoin and Monero are limited by counterparty risk. Trust-minimized insurance would democratize them substantially — the subject of ErgoRich’s Ergo Darkpaper and of ongoing ErgoForum work on Bitcoin on-ramping insurance.
Basis (see Mutual Credit above) is the most direct expression of the thesis: credit created between people, with the chain used only to reinforce trust where it runs out — never to grant permission.
Privacy tooling throughout — the mixing protocol, stealth addresses. Rosen for cross-chaining. The stated priority is useful products for people outside Ergo, or outside crypto entirely, with Ergo as the working horse underneath. Next on the list: more DeFi tooling, concentrated liquidity pools, larger-scale mixers.
Around 2019–20, when the ecosystem was smaller and more centralized, it had roadmaps; they were dropped as a poor fit for a growing community. The proposal now is something roadmap-shaped but different in purpose — a central object for dialogue, for checking progress, and for security and other cross-checks in broad context. Not a plan handed down.
kushti singles out something happening in Ergo developer circles: LLM agents are being used to harden core protocol implementation and ecosystem smart contracts, through constant discussion among humans combined with model cross-checks. A good deal of infrastructure has shipped this way recently.
What comes next
Not everything below runs in production. Some of this is a whitepaper, some a single forum thread, some a prototype without an EIP yet. Ergo has no central team gatekeeping any of it — anyone can propose and ship. Each card says exactly where it stands.
Ordering blocks land roughly every 2 minutes, leaving bandwidth idle in between and wide variance in confirmation time. Sub-blocks — smaller "input blocks" produced roughly once a second — carry transaction data continuously, so a transaction can be felt within seconds while the ordering block still finalizes it. Early write-ups used different ratios (a 2023 blog post illustrated around 128 input blocks per ordering block; 2025 devnet testing ran closer to 60), and the Matrix development branch made that ratio adjustable by miner vote instead of a hardcoded constant. Ergo’s own roadmap calls this active research and development on a devnet, not a mainnet feature.
A sidechain double merge-mined with Bitcoin and Ergo at once, built by BetterMoneyLabs — the team behind ChainCash and Basis — to give a stablecoin- and RWA-focused chain Bitcoin-grade Proof-of-Work security from day one, alongside Ergo’s Sigma contracts. Its design also sketches "dark tokens" that would hide amounts behind homomorphic commitments with stealth-address-like receiver handling — privacy work that exists on paper only so far. Ergo’s ecosystem page lists it “in progress.” The public repository is close to an empty skeleton and the whitepaper’s own tokenomics section is marked unfinished; there is no testnet.
Token-level "policy" contracts that can require whitelist or blacklist checks, mandatory payments, jurisdiction-specific transfer logic, or propagation of the same policy across a transaction’s outputs — enforced by the token’s own script rather than by an EVM-style account contract. kushti proposed it on the Ergo forum (July 2025) as part of Braid’s design, but it isn’t sidechain-exclusive: in the same thread it’s argued the same policies could run directly on Ergo mainnet today, the main obstacle being dapp/wallet support rather than the protocol itself, and kushti later sketched a soft-fork path for it. No EIP and no implementation yet.
From BetterMoneyLabs — the same team building Braid: ChainCash lets anyone issue a spender-signed IOU note that gathers backing as it circulates, and Basis settles the resulting mutual-debt state off-chain, committing only a digest on Ergo. See the Mutual Credit section above for the full picture; contracts and APIs still change often, and some transfer paths need raw Schnorr signatures, so normal wallet support stays limited.
Represents trust as ordinary Ergo boxes: a reputation token carries a positive or negative claim plus optional delegation, so each user builds their own trust graph instead of relying on one global score. Reused across the same circle of projects: Celaut’s distributed service-execution paradigm uses it for node and service trust, a fully on-chain forum and a hash-based file-discovery directory build directly on it, and so do Bene’s own campaign discussions. Game of Prompts goes further still, using reputation proofs and on-chain opinions to vet the judges and creators running its bot competitions. Live on Ergo mainnet, with no EIP and no published staking parameters yet.
Sub-blocks and Braid are Ergo Platform R&D, Global Transfer Policies is a forum discussion whose scope reaches beyond Braid to plain Ergo mainnet, and ChainCash & Basis remain active prototypes — see the Mutual Credit section above for their full status. Reputation proofs is the outlier here: already live on Ergo mainnet today, just without an EIP or published staking parameters yet. Check each linked primary source before building on any of them.
Economics
Mainnet launched on 1 July 2019. ERG had no ICO, no VC allocation and no conventional premine; the Foundation treasury receives 4.43% of total emission through a transparent vesting contract.
Babel fees let a user acquire the ERG needed for miner fees atomically from a liquidity box while paying the provider in another token. The protocol fee is still paid in ERG; no consensus fork was required.
Oracles & bridges
Oracle participants publish independently; the aggregate lands in a single on-chain box.
Any number of contracts can reference that same box in the same block — nobody has to win a race to use it.
Rosen Bridge moves value between chains with watchers and guards doing the coordinating, Ergo as the anchor underneath.
Applications
Ergo’s design is visible in deployed systems — not every proposal is a finished product, and each carries its own smart-contract, liquidity and bridge risks.
Multiple oracle participants publish data and aggregate it on-chain. Data inputs let downstream contracts read a datapoint without consuming the oracle box.
A cross-chain bridge architecture using watchers and guards, with Ergo as its coordination layer. Supported routes and security status should always be checked in the live app.
Ergo’s original decentralized exchange, live since 2021: open-source order-book and AMM contracts that most of the ecosystem’s other exchanges build around rather than each shipping a separate matching engine. Machina Finance is the ecosystem’s exception — it runs its own off-chain grid- and limit-order matching instead of routing through ErgoDex’s shared pools.
A peer-to-peer grid- and limit-order exchange: a trader sets a price range, an off-chain bot triggers the fill, and an on-chain contract verifies and settles it, with no shared liquidity pool for makers to sit in. Pre-launch: the SDK’s first release is a May 2026 alpha, there is no public app or mainnet deployment, and no fee figures are published yet.
A Stability Nexus crowdfunding dApp built on escrow rather than trust: contributors receive a placeholder token, and only receive the project’s real token if the campaign clears its funding goal by the deadline — otherwise the same contract refunds everyone in full. Fully client-side, with a 5% fee charged only on successful campaigns.
Non-custodial privacy protocols based on Sigma proofs. Privacy depends on correct use, anonymity-set conditions and current software status.
Native tokens and box registers support auctions, royalties, metadata and minting without requiring a token contract for every asset.
Decentralized mining-pool infrastructure using Non-Interactive Share Proofs and collateral contracts. Currently testnet software requiring a fully synced node.
From ErgoForum
The community explores designs in public. These are research threads and proposals, not claims that every system is deployed.
Stateless validation, transaction chains, off-chain protocols, sidechains and sub-block research.
How oracle pools and reserve protocols can update participants and parameters without a central operator.
Small ErgoScript contracts can enforce all-or-refund funding conditions directly in the UTXO graph.
Local exchange trading systems explore mutual credit and community accounting rather than speculative token issuance.
Permissionless trade, peaceful counter-economics and censorship-resistant tools form an explicit part of Ergo’s social conversation.
Threshold and distributed signing research builds on the composability of native Sigma propositions.
Why Ergo exists
Ergo’s manifesto frames the platform around permissionless, survivable financial contracts: open participation, resistance to censorship, privacy where needed, and value controlled by users rather than gatekeepers. The aim is not “blockchain for everything”; it is robust contractual money.
Verify, don’t trust
This site separates protocol facts from ecosystem claims, and shipped software from research. Start with the specification and improvement proposals; product availability and parameters can change.