Skip to main content
Testnet: Base Sepolia. Not real money.

Risk Disclosure

Last updated: May 24, 2026

Real money. You can lose all of it. No recourse.

When Djinn is on Base mainnet, every dollar you deposit is real USDC. Bugs, exploits, validator failures, key loss, phishing, and operator mistakes can permanently destroy your funds. There is no insurance fund, no chargeback, no customer-support recovery path, and no one to sue for restitution. The protocol is decentralized and the team disclaims liability to the maximum extent permitted by law.

During soft launch, per-trade purchases are capped at $1 as a defense-in-depth backstop. As the cap is raised over time, keep using small amounts you can afford to lose. The cap is a ceiling, not a recommendation.

If you cannot read the contract source code yourself at github.com/djinn-inc/djinn and verify what the UI is asking you to sign, you are not the intended audience for Djinn yet.

This Risk Disclosure is incorporated into the Terms of Service. It supplements, but does not replace, the risks described there. By using Djinn, you acknowledge that you have read, understood, and accepted the risks below.

1. Network Status & Soft Launch

Djinn is currently transitioning from Base Sepolia (a public test network, tokens have no cash value) to Base mainnet (real USDC, real economic value). The active network is shown in the persistent banner at the top of every page. While the banner reads “Testnet,” no real funds are at risk. While the banner reads “Soft Launch,” you are sending real USDC to real smart contracts and the entire content of this page applies in full.

Soft launch bounds. The initial mainnet configuration caps single purchases at $1 USDC. This cap exists because we expect to find bugs early, and the cap bounds how much you can lose to any single broken interaction. As we accumulate uneventful operating time, the cap will be raised in steps governed by the on-chain 72-hour timelock. You should not treat a higher cap as an endorsement to deposit that amount. Use only what you can afford to lose entirely.

No backwards-compatibility for testnet positions. Signals, balances, and history on Base Sepolia do not carry over to mainnet. Sepolia data remains on chain indefinitely but is not used by the mainnet protocol or front-ends.

2. Smart Contract Risk

Djinn is a set of smart contracts on the Base blockchain. Smart contracts are software. Software has bugs. An undiscovered bug, a flawed upgrade, or a compromise of the governance keys could cause you to lose all funds you have deposited. The contracts are open source and have been reviewed internally and by third parties. That does not mean they are free of defects. No review finds every bug.

The Djinn contracts are upgradeable. A privileged multisig operated by Djinn Inc., acting through a 72-hour on-chain timelock on mainnet, can deploy new implementations. An attacker who compromises the multisig, or Djinn Inc. itself acting in bad faith, could theoretically push a malicious upgrade. The timelock gives you time to exit, but only if you are watching the chain.

3. Signal Quality Risk

Signals are analytical predictions. They can be wrong. The service-level agreement mechanism (SLA) provides partial structured compensation if a Genius's aggregated Quality Score is negative, but it does not guarantee accuracy. You can buy signals that turn out to be worthless, and in some scenarios (low-volume Geniuses, settlement delays, protocol bugs) you may receive less compensation than the SLA formula implies.

Past performance shown on any Genius's track record does not predict future performance. A Genius with a long winning streak can start losing tomorrow. Track records on Djinn are verifiable on-chain, but verifiability is not the same as skill.

What Quality Score is and is not. Quality Score (QS) is the on-chain accounting primitive that determines whether a Genius's posted collateral is slashed and, if so, by how much, at the end of an audit batch. QS is computed deterministically from on-chain purchase records and the validator network's publicly-verifiable event-outcome feed. QS is not a wagering outcome — Djinn does not take a house position, does not accept side bets, and pays out only from collateral already posted by the seller before any purchase occurred. QS is also not investment advice, not a forecast of any individual signal's future outcome, and not a recommendation by Djinn Inc. to purchase from any particular Genius. It is a backward- looking dollar measure of what the protocol has already settled.

4. Validator and MPC Risk

The Djinn Protocol depends on a validator network registered on Bittensor Subnet 103. Validators hold Shamir shares of master seeds, perform secure multi-party computation (MPC) to derive signal keys on purchase, and vote on audit outcomes. If too many validators go offline simultaneously, signal purchases can fail, decryption can fail, and audits can stall. Settlement (collateral slashing or release) depends on a validator quorum reaching consensus on on-chain outcomes. Quorum failures, censorship by a colluding subset, or incentive misalignment can delay or prevent settlement.

The validator set is bounded, not unbounded. Subnet 103 has a finite number of validator slots (currently in the low tens), and the slot set is governed by Bittensor's native registration and consensus mechanisms. Slot operation is permissionless to enter via Bittensor registration, but at any given moment the operating set is a small, identifiable group whose hotkeys and operating addresses are public on chain. This means that (a) the “decentralization” property of Djinn is finite — a coordinated action by a sufficient fraction of operators could materially affect settlement, and (b) the identities (or at least hotkeys) of the operators are visible to anyone who queries the Bittensor metagraph or Djinn's /network page. Djinn Inc. or its affiliates may themselves operate one or more slots; see Terms §6b for the affiliate-participation disclosure.

The MPC protocol itself is a newly deployed cryptographic system. A cryptographic flaw in the MPC implementation could, in principle, allow an attacker with enough compromised validators to reconstruct a seed and read signals, or to sign fraudulent settlements. Current bootstrap parameters (Shamir t-of-n) are documented in the whitepaper and are designed to tolerate a minority of bad validators; they do not tolerate majority collusion.

5. Bittensor Network Risk

Djinn validators and miners are registered on Bittensor Subnet 103 and are rewarded in TAO, the Bittensor native token. The economic security of the validator set depends on TAO having and retaining value, on the Bittensor governance process not materially changing subnet parameters in a way that breaks Djinn, and on the subnet continuing to be registered in good standing. A deregistration event, a hostile governance proposal, or a severe TAO price collapse could weaken validator incentives to the point that settlement is unreliable. TAO holders are subject to their own regulatory, custody, and market risks, independent of Djinn.

6. Blockchain Risk

Transactions on the Base blockchain are irreversible. If you send funds to the wrong address, or sign a malicious transaction prompted by a phishing site impersonating Djinn, those funds are unrecoverable. Network congestion can cause transaction fees to spike. A Base outage, a sequencer failure, or a reorg deeper than the finality threshold could delay, re-order, or, in edge cases, invalidate transactions you believed were confirmed. Base is operated by Coinbase and has its own disclosures and risks; consult Coinbase's materials for details on Base's architecture and governance.

7. Stablecoin Risk

The stablecoin used for settlement is USDC, issued by Circle. USDC is a private stablecoin; it is not a bank deposit, is not FDIC-insured, and is not a claim on the United States government. Its one-to-one peg with the U.S. dollar depends on Circle's ability to honor redemptions. A loss of the peg (de-pegging), a regulatory action against Circle, or an operational failure at Circle could cause USDC to trade below par, which would reduce the real value of your Djinn balance.

8. Regulatory Risk

The legal status of information marketplaces, cryptocurrencies, smart contracts, decentralized finance, and related technologies varies by jurisdiction and is evolving rapidly. Regulators may, at any time, characterize parts of Djinn as a regulated activity (for example, as gambling, money transmission, a securities offering, or commodity trading), whether or not we consider that characterization correct. Regulatory action could force Djinn to restrict service in your jurisdiction, shut down user-facing websites, freeze features, or comply with information requests about users within the reach of applicable law.

U.S. state sports-wagering risk. Sports wagering is regulated state-by-state in the U.S. Several states prohibit sports wagering entirely (for example, Utah and Hawaii); others criminalize unlicensed online sports wagering (for example, Washington); and many permit it only through state-licensed operators. Djinn does not place, accept, or process wagers, and the protocol is structurally incapable of doing so (see Terms §2). However, if you use a signal you purchase on Djinn to place a wager that is unlawful in your jurisdiction, you are the one engaging in the unlawful conduct, not Djinn — but the existence of unlawful downstream use can still trigger investigation of Djinn and reduce your own access to the service. If you are unsure whether sports wagering is lawful where you are, consult a lawyer before purchasing any signal.

Possible recharacterization as a securities offering. A Genius posts collateral, sells information, and shares in the economic outcome of a buyer's use of that information through the Quality-Score-driven slashing mechanism. Although our position is that this structure is a collateralized information sale rather than an investment contract — each Genius-Idiot transaction is bilateral, no pooling of buyer capital occurs into a common enterprise, the buyer is paying for information rather than for a stake in the seller's future actions, and the seller's collateral is bounded and pre-posted rather than recursively reinvested — we acknowledge that a regulator could disagree. If a regulator with jurisdiction over you determines that some aspect of Djinn constitutes a securities offering, money transmission, gambling, or other regulated activity, Djinn may be required to restrict or terminate service in your jurisdiction, and you may face independent personal consequences. You accept this risk by using the protocol.

Your use of Djinn may itself be illegal in your jurisdiction. You are solely responsible for compliance with all laws applicable to you, including gambling prohibitions, securities laws, money-transmission rules, tax reporting, and sanctions law. Djinn does not offer legal advice and is not a substitute for consulting your own counsel.

9. Custody and Private Key Risk

Djinn is non-custodial. Your wallet private key, and if you use a smart wallet your recovery method, are solely under your control. If you lose your key, your seed phrase, or access to your wallet provider, no one at Djinn can restore your funds. If your key is stolen (through a phishing attack, device compromise, malicious browser extension, or other means), the thief can take everything you have deposited. We will never ask for your private key, seed phrase, or password for any reason.

10. Front-End and Infrastructure Risk

The djinn.gg website is one front-end to the protocol. If the front-end is compromised (DNS hijack, TLS compromise, supply-chain attack on our hosting or dependencies), you could be served malicious code that prompts you to sign transactions that drain your funds. To mitigate this, advanced users can interact directly with the smart contracts from a wallet of their choice, or from an independent front-end. The contracts are permissionless; the website is not the protocol.

11. Counterparty Risk (Geniuses and Idiots)

Djinn reduces, but does not eliminate, counterparty risk. Collateral locks create structural protection for Idiots if Geniuses underperform, and Idiot balances are held in contracts, not by Geniuses. But slashing is capped at posted collateral. A Genius who is catastrophically wrong on a long series of signals will only refund up to the collateral they posted; further losses attributed to their signals are not covered. Likewise, an Idiot who defaults on platform fees can only be pursued to the extent of their on-platform balance.

12. Market Manipulation and Insider Risk

Djinn cannot prevent all forms of manipulation or insider abuse. A sophisticated actor could attempt to wash-trade their own signals to inflate a track record, collude with other participants to steer audits, or exploit information asymmetries (advance injury reports, line-movement arbitrage, inside access to officiating or lineups) in ways that disadvantage other users. Our Acceptable Use Policy prohibits these behaviors, but prohibition is not the same as prevention. Bear this risk in mind when evaluating any track record.

13. Operational Risk

Djinn is operated by a small team. The team can make mistakes. Configuration errors, botched deployments, stuck cron jobs, validators running stale versions, or an outage of a third-party service (odds provider, RPC provider, hosting) can cause transient or persistent degradation of service. In rare cases, operational mistakes could affect settlement or pricing. Monitoring and incident response are in place but are not infallible.

14. Credit and Recovery Risk

Djinn Credits are non-transferable, non-cashable platform credits that function only as a discount on future signal purchases. They are not a debt instrument and are not redeemable for cash. A protocol change, contract upgrade, or discontinuation of the service could render credits unusable. Treat them as store credit with all that implies.

15. Loss of Service

Djinn may cease to operate. The team may dissolve, the company may enter bankruptcy, a regulator may compel takedown of the website, or the Bittensor subnet may be deregistered. In any of these events, the smart contracts remain on-chain and, to the extent your funds are in them and withdrawal functions remain callable, you can withdraw. But a fully operational ongoing service is not guaranteed, and the user-interface path to withdrawal may be unavailable.

16. No Financial Advice

Nothing published by Djinn, its officers, employees, contractors, or community members is financial advice, investment advice, legal advice, tax advice, or a recommendation to wager on any outcome. Any decisions you make based on information obtained through Djinn are your own.

17. No Warranty

Djinn is provided “as is” and “as available” without any warranty, express or implied. Nothing on this page, in the whitepaper, on the website, or in any other Djinn material constitutes a warranty of functionality, uptime, accuracy, fitness for purpose, merchantability, non-infringement, or freedom from defect.

18. No Recourse and Do Your Own Research

Djinn is a decentralized protocol. There is no recourse if your funds are lost — no chargeback, no insurance fund, no central customer-support team that can reverse a transaction. The cofounders, operators, validators, and miners disclaim all liability to the maximum extent permitted by law for any loss arising from your use of the protocol, including but not limited to losses from smart-contract bugs, cryptographic flaws, validator outages or collusion, front-end compromise, dependency compromise, signal-quality variance, oracle misreads, regulatory action, custody errors on your side, phishing, bridge failures, RPC failures, or any combination of the above.

We make no guarantees of accuracy, uptime, profitability, recovery, timeliness of settlement, or continuity of service. We do not underwrite signal performance. We do not vouch for any individual Genius or Idiot. Aggregated track records visible in the UI are unaudited statistical summaries of on-chain events, not investment recommendations.

Do your own research. Read the contract source code at github.com/djinn-inc/djinn. Read the whitepaper. Read the on-chain logs. Verify the contract addresses shown in the UI against block explorers and against multiple independent sources before signing any transaction. If the UI says one thing and the contract says another, trust the contract.

Start small. Stay small. Soft launch caps every purchase at $1. When caps rise, keep your own commitments small until you have personally watched the protocol behave correctly with your own funds across many cycles. Do not deposit any amount you could not immediately accept losing.

19. Acknowledgment

By connecting a wallet to Djinn, interacting with the API, or depositing funds, you acknowledge that you have read this Risk Disclosure in full, that you understand each risk described, that you accept full responsibility for your own conduct, and that you waive all claims against the cofounders, operators, validators, miners, and any affiliated entities for any loss that may arise from your use of the protocol, regardless of cause or fault.