What to use instead of Augur

The product

Augur is the oldest thing in this category. It held one of Ethereum's first crowdsales in 2015 and launched on mainnet in 2018 — those two dates are the project's own, from its blog post of 21 July 2026 — and it was built on a position that was unusual then and is still rare: no company and no committee gets the final word on what happened. A market named a reporter who posted the first outcome after close; anyone who disagreed could stake REP against it; each round of disagreement cost more than the last; and if the dispute escalated past every round, the protocol itself split.

That last step is the reason this page exists in 2026 rather than 2021.

The mechanism ran. On 8 April 2026, a community member — Micah Zoltu, who had crowdfunded the attempt the previous August — pushed a dispute over whether the Artemis II mission launched successfully far enough to trigger a fork. An escalation game ran from April into early June. Then the protocol did what its whitepaper says it does: it created one child universe per possible outcome, and every REP holder had 60 days to move their tokens into the one they believed was right. The window ran from 4 June 2026 at 01:00 UTC to 3 August 2026 at 01:00 UTC.

The numbers are on chain and we checked them against the announcement rather than taking it on trust. Of the original 11,000,000 REP, 6,547,546.66 migrated — about 59.5 percent. The REPv2_Yes_1 token at 0xCf6A0A7826fa124B7705d6f3c675eAD76f1e540D reports a total supply of 6,545,760.433666705616389974 and the REPv2_No_1 token at 0x2F4005456c2F098358213f01DbE34abDAa2989A4 reports 1,786.2274008665274. Both figures match the announcement to the decimal. The pre-fork REPv2 contract still reports 2,697,575.656086505648896712 left behind in the parent universe, and the 2015-era REPv1 contract still reports its original 11,000,000. The parent universe and both children all hold code and all still respond.

So the protocol is not dead, and a page that called it dead would be wrong. What does not exist is an application. The project's own FAQ, read on 19 September 2026, puts it plainly: not as a new Augur-native prediction-market application, because one has not yet relaunched, and the site is not a trading interface. Everything www.augur.net serves today is a reboot site — a mission page, a blog, a whitepaper library, a learn section and a link to v3.augur.net, which is ForkWatch, a read-only archive of the migration. app.augur.net, docs.augur.net and blog.augur.net do not resolve at all.

The on-chain record says the same thing more quietly. In the 40,000 blocks to 19 September 2026 — roughly five and a half days — the Augur v2 registry contract at 0x23916a8F5C3846e3100e5f587FF14F3098722F5d emitted 1,101 events, and they are almost entirely bookkeeping: 731 REP balance changes, 365 token transfers, one mint, one reporting-fee recalculation, one dispute window opened on schedule, one share-token balance change, and one address claiming the proceeds of a settled position. No market was created. No outcome was reported. No dispute was contributed to. Of the 731 balance changes, 605 were in the parent universe — the one the project says is expected to be worthless — and 126 in the surviving one; we can see that they happened and we cannot tell you what they were.

This is why the header on this page says the terms changed rather than shut down. Every other subject in this section stopped. Augur did not: it is between versions, with the old one intact and unusable and the new one not deployed. A foundation, Lituus, has funded two teams since March 2025. One works with ChainSafe on a redesigned oracle, in a repository last touched on 17 September 2026. The other, Dark Florist, is building a successor protocol and three interfaces in the open — a repository called zoltar, under the Unlicense, whose most recent commit on the day we checked was that same day. It carries a file of mainnet deployment addresses; we looked up all fifteen, and fourteen hold no code on Ethereum mainnet. The successor is real, it is public, and it is not deployed.

Meanwhile the token trades. Kraken lists the migrated token as AUGUR and its AUGUR/USD pair was online on 19 September 2026, last trade $0.8546, with 1,228 tokens changing hands over the previous 24 hours. On Uniswap the largest pool held about $158,000 and did about $222 in the same day. CoinGecko renamed its old entry to "Augur [OLD]" and still prints a price for the stranded token — around $1.05, above the live one, on volume of a few hundred dollars from one Indonesian exchange and two old Uniswap pairs. A price on a token nobody can use is not a market; it is a listing nobody has got round to removing.

What to use instead

Augur was four things in one protocol — a no-permission market factory, a self-custody venue, a dispute system nobody owned, and a codebase you could read instead of a terms page. They do not land in the same place.

If what you had was a wallet and no identity check, Polymarket is the closest working venue and the closest in design. An account is a wallet, there is no identity check in the ordinary flow, and outcomes are proposed to the UMA optimistic oracle with a bond behind them and a two-hour challenge window; a proposal disputed twice escalates to a vote of UMA token holders. That is the only resolution process left in this catalogue that a third party can enter with their own money, which was the whole of Augur's argument. It is not the same argument: the voters hold a different token, there is no universe split behind them, and Polymarket blocks the United States outright. Limitless is the same accessibility with a plainer mechanism — YES/NO shares fully collateralised in USDC on Base, an order book rather than a curve, 1.5 percent to the taker, no identity check.

If what you had was the habit of opening your own markets, Limitless is the only pick here where you can do that with real collateral and take a creator fee on it, which is the closest thing to Augur's market-creation bond and creator-fee pair. Manifold is the unrestricted version and comes with the obvious caveat: anyone opens a question, the creator resolves it, the code is MIT-licensed and readable, and the currency buys nothing. If you were creating markets to find out what people thought, that is a small loss. If you were creating them to earn the fee, it is the whole thing.

If what you want is the mechanism Augur used, Myriad prices with an AMM across four chains, settles in stablecoins, needs a wallet and nothing else, and takes 0.5 percent — the cheapest here. Augur v2's own liquidity was pooled the same way. What Myriad does not have is Augur's back half: the team resolves every market and there is no challenge path.

If you simply want a venue with a lot of questions that will still be there next year, Futuur runs the broadest subject list in this catalogue, in USDC or USDT alongside a play-money book on the same questions, with no identity check and an editorial team that names the public source each market resolves against.

If you are in the United States, none of the above is open to you, and Augur effectively never was either — there is no US-regulated venue in this catalogue that resolves the way Augur does, because resolution by a token-holder vote is not something a designated contract market does. The regulated answers are in the prediction market venues category, and they are a different product with a different bargain: dollars, an identity check, and a rulebook filed with a regulator instead of a fork.

And if what you actually wanted was a protocol rather than a product, nothing here replaces it, and it is worth saying so rather than pretending. Four of the five picks are hosted businesses with terms pages. The fifth pays in play money.

What you give up

A market you can open without asking anyone. Augur's Universe contract exposes createYesNoMarket, createCategoricalMarket and createScalarMarket to any address willing to post the bonds, with the creator naming the fee and the reporter. Limitless has creator markets with a creator fee and is the nearest working equivalent; Manifold is unlimited and pays in mana. Everywhere else in this catalogue an operator decides which questions exist, and a question nobody at the venue wants is a question that does not exist.

Scalar and categorical markets. Every venue in the list above is binary YES/NO. Augur v2 had markets with up to eight named outcomes and scalar markets whose payout is a number in a range rather than one of two answers. If those are what you were using, you are not changing venue, you are changing what you can express — and the honest first step is to work out whether the question you care about can survive being rewritten as yes or no.

Resolution that nobody can overturn, and the fork behind it. This is the real loss and 2026 is the year it became demonstrable rather than theoretical. On Myriad and Futuur a team resolves and there is no challenge mechanism at all. On Limitless a price feed resolves the automated markets and the team resolves the rest. Polymarket's UMA path is a genuine dispute process a stranger can join, and it is still a bonded proposal escalating to a token vote — not a stake-weighted escalation game ending in the protocol splitting in two and every holder having to choose. Augur ran that to completion this year. Nothing you can trade today implements it.

Rules compiled rather than published. The v2 contracts are MIT-licensed and public, which means the bond sizes, the fee caps, the dispute rounds and the fork threshold were constants you could look up and disagree with. What the replacements charge and how they resolve is a claim on a page, and a page can be edited between your reading it and your position settling.

What you get back is that there is something to trade. Every venue above is currently matching orders, taking deposits and paying out. That sounds like faint praise until you have spent an afternoon confirming that a protocol works perfectly and has no markets in it.

Migration notes

If you did not migrate by 3 August 2026, the tokens are still yours and the door is shut. The deadline lives in the fork contract, not in a policy, which means there is no support address that can make an exception and no governance vote that can reopen it. Your REPv1 or REPv2 still sits in your wallet and still transfers. The project says it expects it to be worthless. Read every offer of late migration, recovery or a replacement claim as an attempt to take what is left, because both the project and its own ForkWatch page say exactly that, unprompted.

Check the contract address before you touch anything called REP. There are now four live tokens with nearly the same name: REPv1 from 2015, REPv2 from the v2 migration, and one child token per fork outcome. Exchanges do not use the on-chain symbols — Kraken lists the surviving token as AUGUR — so the address is the only thing that disambiguates them. The one development continues on is 0xCf6A0A7826fa124B7705d6f3c675eAD76f1e540D. A wallet that shows you nothing is usually just a wallet that has not added the new token yet; 0x2F4005456c2F098358213f01DbE34abDAa2989A4 is the losing side and holds 1,786 tokens in total.

Where it was custodied decided it for you. ForkWatch's archived record lists Kraken, OKX and Bitpanda as supporting the fork and Coinbase, Gate.io and Upbit as not. Kraken's notice, last updated 17 July 2026, describes a 1:1 migration of custodial balances into the surviving universe with no client action required, trading paused on 24 July and AUGUR credited by the end of 31 July — and adds that clients in the EEA and Canada could withdraw the new token but not trade or deposit it, pending a MiCA whitepaper filing. On 19 September 2026 Coinbase's public API still described REP-USD as delisted with trading disabled, against the pre-2020 token contract. If you held REP somewhere and have never seen a new balance, that exchange is the only party who can tell you what happened to it.

Anything you built that reads Augur is reading a fork it does not know about. A dashboard, a bot or a notebook pointed at the v2 contracts will keep working and keep returning numbers, because the contracts are live. What changed underneath is which universe those numbers describe and which token address is the current one. Nothing errors. Make the token address an assertion at startup rather than a constant somebody set in 2021.

Do not follow the old documentation. The v2 repositories are not archived, which reads as health and is not: the client's last published release is v2.1.13 from 5 March 2021 and the last commit on the main repository is from 23 July 2023. The GitBook help centre still returns 200 and still explains how to sign up, add funds, trade, read a portfolio and create markets, and still presents Augur Turbo on Polygon as a live product. Treat all of it as history.

Nothing transfers to any of the venues above. There is no snapshot, no partner venue and no recovery token. Every pick here is a new account on a different chain funded with different collateral, and the contracts are not the same contracts. Anyone telling you otherwise is selling you the replacement.

Re-check before you act on this page. This is the one subject in this section where the direction of travel is towards working rather than away from it. A funded team is writing the successor protocol in public and has a mainnet address file ready for contracts it has not deployed yet. Everything above describes 19 September 2026; if you are reading it much later, the first thing to check is whether those addresses have code.