Anoma is an intent-centric Layer 1 blockchain protocol where users express what they want as an outcome rather than specifying the exact sequence of transactions needed to get there. Instead of manually approving tokens, choosing routes, and submitting transactions, a user signs an intent, a cryptographically committed statement of desired state change. Anoma's counterparty discovery and settlement layer then resolves that intent, potentially matching it with other users' intents, without the user managing execution details. As of Q1 2026, Anoma remains in research and testnet phases; its mainnet has not launched.
​

The project was founded by the Heliax research lab and has published a formal specification for its Resource Machine model, which replaces the account-based and UTXO-based state models used by Ethereum and Bitcoin respectively. Namada, a sibling privacy-focused chain, shares Anoma's core architecture and is further along in deployment.
​

Intent-centric architecture is a blockchain design where users declare a desired outcome ("I want to hold USDC worth $500, currently holding ETH"), and the protocol handles how that outcome is reached. The user's signed intent is broadcast to a network of solvers who compete to find valid execution paths. This shifts execution complexity from user wallets to a competitive solver market.
​

Traditional blockchains like Ethereum require users or wallets to construct discrete transactions: approve a token, call a swap function on Uniswap, submit the encoded calldata, and pay gas at each step. When multiple steps span chains, the complexity compounds. An intent-centric model collapses this into a single declarative statement. The user specifies preconditions (what they have) and postconditions (what they want) and signs a commitment to that delta. Solvers, which can be searchers, market makers, or specialized bots, compete to satisfy the intent and claim any available surplus.
​

The Anoma whitepaper formalizes intents as partial transactions that become valid only when combined with complementary partial transactions from other parties. A swap intent from Alice and a swap intent from Bob, if they are counterparties, compose into one complete atomic transaction. Neither executes unless both do, removing counterparty risk from the user-facing layer.
​

Intent-based bridges like Across Protocol apply a narrower version of this model to cross-chain transfers: users declare a desired destination state, relayers advance the funds, and the protocol reimburses relayers via a Hub chain. Anoma's model is more general: intents can describe any state transformation across any resource type, not just token transfers.
​

The Anoma Resource Machine (ARM) is a unified execution model that replaces both the account and UTXO models. Resources are typed objects with defined validity predicates. A transaction is valid when every predicate passes and the overall balance of resources zeroes out, meaning nothing is created or destroyed without authorization.
​

In Ethereum's account model, a "balance" is a number stored in a mapping inside a contract. To transfer tokens, a transaction modifies that mapping. Composability between contracts requires explicit function calls and shared interfaces. In Bitcoin's UTXO model, value exists as discrete spent outputs; a transaction consumes UTXOs and creates new ones. Neither model natively expresses partial transactions that match with counterparties.
​

The ARM represents every asset, vote, NFT, or data record as a typed resource with three components:

When a user submits an intent, they create a partial transaction that specifies consumed resources (what they give) and produced resources (what they receive) but leaves the counterparty's half unspecified. Solvers complete the partial transaction by supplying the matching half. The ARM verifies that the combined transaction balances to zero across all resources and that all validity predicates pass, then commits it atomically.
​

This architecture makes programmable settlement a built-in feature. A validity predicate could enforce that a resource can only be transferred during a specific time window, or only to addresses that satisfy a privacy proof, or only as part of a multi-party agreement. The Anoma GitHub repository contains the reference implementation in Rust, with the ARM formalized in the Juvix functional language.
​

Typhon is Anoma's proposed BFT consensus protocol, designed to order transactions submitted by solvers without requiring those transactions to be fully specified upfront. Typhon separates mempool ordering from execution validity, allowing the network to reach consensus on the ordering of intent-resolution attempts before verifying whether the ARM confirms the final state. This enables parallelism in intent processing that single-threaded execution environments cannot achieve.
​

Standard BFT consensus protocols like Tendermint, used by Cosmos chains, require full transaction validity before a block is finalized. Every validator re-executes every transaction. Typhon's design, as documented in the Heliax research blog, instead treats intent bundles as partially specified objects. Validators agree on ordering; the ARM then determines which bundles are valid state transitions. Invalid bundles are dropped post-ordering without forking the chain.
​

As of Q1 2026, Typhon has not been deployed in a production environment. Heliax has published formal specifications and early testnet results, but the protocol is still undergoing research-phase iteration. The team has noted that Typhon's parallelism properties become relevant at high solver throughput, which the testnet does not yet stress-test at scale.
​

Anoma's broader consensus model is described in the Anoma specification site, which separates consensus (Typhon), state (Resource Machine), and the intent gossip layer (the Interpool) into distinct protocol concerns.
​

Counterparty discovery in Anoma is handled by the Interpool, a peer-to-peer gossip network where nodes broadcast partial transactions. Solvers monitor the Interpool, pattern-match complementary intents, construct complete transactions, and submit them to the consensus layer. No central matching engine exists; discovery is fully decentralized and permissionless, allowing any party to operate a solver.
​

When Alice signs an intent to exchange ETH for USDC, her wallet client submits the partial transaction to one or more Interpool nodes. Those nodes gossip the intent across the network. A solver running a pattern-matching engine sees Alice's partial transaction and, if Bob has a complementary intent to exchange USDC for ETH, the solver bundles both halves into a complete transaction. If no direct counterparty exists, the solver can introduce intermediate steps (routing through a liquidity pool or bridging across chains), as long as the overall transaction satisfies both parties' validity predicates.
​

This design gives solvers flexibility in how they fill intents. A solver could match Alice directly with Bob (peer-to-peer settlement), could route through an automated market maker like Uniswap, or could use cross-chain liquidity. From Alice's perspective, the outcome is the same: she receives what she declared she wanted, or the transaction does not execute. The solver competition for surplus is what drives efficient pricing.
​

The model differs structurally from Across Protocol's approach to intents. Across uses a centralized Hub chain and a whitelist of relayers with required bonds. Anoma's Interpool is permissionless at the solver layer and generalizes beyond asset transfers to any resource type the ARM can represent. The Anoma blog post on the Interpool design details the gossip protocol and solver interface.
​

Traditional blockchains require users to construct the complete execution path of every transaction. Anoma inverts this: users submit underspecified intents and the network resolves execution. Users express desired outcomes, solvers compete on efficiency, and the Resource Machine enforces correctness through validity predicates rather than user-supplied calldata.
​

The table below summarizes the structural differences between Anoma's design and the two dominant blockchain execution models.
​

One consequence of the ARM model is that privacy becomes a per-resource property rather than a chain-level property. A resource's validity predicate can require a zero-knowledge proof of ownership without revealing the owner. Namada, the sibling chain that uses Anoma's Resource Machine with MASP (Multi-Asset Shielded Pool) proofs, demonstrates this: users can transact privately across multiple token types on a single chain because privacy is enforced at the resource level, not via a mixer or L2 escape hatch.
​

Anoma's ecosystem centers on two primary projects: the Anoma protocol itself (the research-stage L1) and Namada (a production-track privacy chain sharing the ARM and intent architecture). Additional projects build tooling, developer interfaces, and applications around these two chains, with most activity concentrated in Q1 2026 on testnet environments rather than mainnet deployments.
​

Namada is a sovereign, proof-of-stake blockchain built by Heliax as the first practical deployment of the Anoma Resource Machine. Its key addition is the MASP circuit, which enables private multi-asset transfers: a user can shield ETH, USDC, and DAI into a single shielded pool and transact between them privately. Namada shipped its mainnet in Q1 2024 and has been the primary live demonstration of Anoma's architectural concepts. Namada's documentation describes the shielded set and MASP integration in detail.
​

The Anoma L1 testnet, as of Q1 2026, allows developers to submit intents using the Juvix language and test solver logic against the ARM. Heliax provides a Juvix compiler, a local node simulator, and a test Interpool environment. The project's GitHub activity shows active development on the ARM specification and Typhon consensus implementation, with no public mainnet launch date announced as of Q1 2026.
​

Several teams have begun building solver infrastructure targeting Anoma's Interpool, though most are in pre-production stages. The intent-matching problem Anoma poses, finding complementary partial transactions across a decentralized gossip network, is algorithmically richer than the mempool-scanning approach used by MEV bots on Ethereum, and has attracted research interest from academic groups studying combinatorial auctions and mechanism design.
​

Intent-based systems exist across a spectrum from narrow (Across Protocol's cross-chain transfer intents) to general (Anoma's full-state-transformation intents). Anoma's research-stage status means that live intent-based systems with deployed mainnet infrastructure currently handle more real-world volume, while Anoma offers a more complete theoretical model for programmable settlement.
​

The table below compares four intent-relevant systems across key dimensions.
​

Anoma's "Intents Aren't Real" post articulates why the team considers current "intent-based" systems as partial implementations: they accept user-facing simplicity but still require a known execution environment (a specific DEX, a specific bridge) rather than fully generalizing over all possible execution paths. Anoma's Resource Machine is designed to generalize completely, at the cost of requiring a new consensus and execution layer.
​

Anoma's architecture points toward cross-chain payment flows where a user signs a single intent specifying a desired outcome ("pay 100 USDC on Base from whatever I hold on Ethereum"), and solvers handle bridging, swapping, and routing. As of Q1 2026 this is theoretical; no production payment flows run on Anoma mainnet, which has not launched.
​

Eco's routing infrastructure currently handles the practical side of this problem: abstracting chain selection, liquidity routing, and settlement for stablecoin transfers across 15+ supported chains. As intent-based models mature on live networks, the solver and routing layers that projects like Eco operate will increasingly implement intent semantics, accepting outcome specifications and handling execution complexity internally rather than surfacing it to end users or developers.
​

As of Q1 2026, Anoma's L1 is in testnet phase only. No production mainnet has launched. Namada, a sibling chain sharing Anoma's Resource Machine and intent architecture, is live on mainnet as of Q1 2024. Developers can interact with Anoma's ARM and Interpool through testnet environments provided by Heliax.
​

Namada is a production-deployed, privacy-focused blockchain that uses Anoma's Resource Machine and Multi-Asset Shielded Pool for private multi-token transfers. Anoma is the broader research-stage L1 protocol. Namada demonstrates Anoma's core architecture in a live environment, while Anoma itself continues pre-mainnet development.
​

Across and UniswapX support intents scoped to specific operations: cross-chain transfers (Across) and token swaps (UniswapX). Anoma's intent model is general: an intent can describe any state transformation across any resource type. This generality requires a new execution model (the Resource Machine) rather than retrofitting intents onto existing smart contract platforms.
​

Anoma's smart contract and intent logic is written in Juvix, a functional programming language developed by Heliax for the ARM's validity predicate system. Juvix compiles to ARM bytecode. Standard Solidity or Move contracts do not run directly on Anoma without transpilation, reflecting the fundamental difference between Anoma's Resource Machine and EVM-based execution.
​

The Interpool is Anoma's peer-to-peer gossip network for broadcasting and discovering partial transactions. Solvers connect to the Interpool, monitor incoming intents, and construct complete transactions by matching complementary partial transactions. Unlike a centralized order book, the Interpool has no single matching engine. Any solver can participate without permission or bond.
​

Sources and methodology. Protocol architecture details sourced from the Anoma whitepaper, Anoma specification site, and Heliax research blog. Mainnet status for comparison protocols verified against official documentation as of Q1 2026. Stablecoin market data from DeFiLlama, April 29, 2026. Figures refresh quarterly.
​

By Eco research. Updated Apr 2026.
​