[2611 Grant] BitChill: V2 audit and relaunch of non-custodial Bitcoin DCA

The ask

USD 20,000, Strategic tier, paid in rBTC across four key milestones:

  • Milestone 1: USD 4,000 (20%)
    Paid after approval and execution of the M1 on-chain proposal. Covers locking the V2 code at a public version (tag), signing the Zealynx Security engagement, and publishing the audit scope and the USDRIF/USDT0 price policy.
  • Milestone 2: USD 5,500 (27.5%)
    Paid once the Zealynx audit is delivered and its fixes have been reviewed.
  • Milestone 3: USD 4,500 (22.5%)
    Paid when V2 is live on Rootstock mainnet.
  • Milestone 4: USD 6,000 (30%)
    Paid after the day-90 impact review. The amount depends on usage: nothing below USD 3,500 of volume, and the full amount at USD 10,000 or more.

What it pays for:

  • The V2 audit. BitChill’s 50% share of the Zealynx audit, passed through at cost: USD 3,000.
  • Fixes for both audits (Kann and Zealynx): USD 3,000.
  • Relaunch on mainnet. Deployment, LayerBank integration and monitoring, and a withdrawal page for V1 users: USD 3,000.
  • Fair pricing for USDRIF and USDT0. Off-chain controls that keep each purchase within a published price standard: USD 4,500.
  • Reporting and community. Monthly reports, AMAs and user support: USD 2,250.
  • 12 months of maintenance: USD 3,750.
  • 12 months of running costs. Keeper gas, server, RPC and domain: USD 500.

Summary

Project: BitChill

Classification: Growth stage (live on mainnet since September 2025) · DeFi / consumer savings · Rootstock users who want to accumulate rBTC from stablecoins without timing the market.

Submission date: 13/11/2026

BitChill is a non-custodial bitcoin DCA protocol on Rootstock. Users deposit a stablecoin, choose an amount and a frequency, and a keeper buys rBTC for them on schedule. Between purchases the stablecoins sit in a Rootstock lending market and earn interest. Only the user can withdraw their stablecoins, yield or rBTC. V2 also offers idle schedules: the stablecoins are not lent, so they earn no yield but carry no lending-venue risk.

V1 has been in production for a year. In that time users spent close to 60,000 DOC on scheduled purchases and received close to 0.8 rBTC, across about 1,000 purchases in 90+ batches. DcaManager, the entry-point contract, processed 717 successful transactions: 617 from 73 user wallets, 99 from the keeper and 1 admin. Deposited principal peaked at close to USD 26,000 (25,952.30 at block 8,461,016, on 26 January 2026 at 06:54:25 UTC). Since April the venues V1 relied on have failed one after another: Tropykus announced its shutdown, Sovryn’s DOC pool became illiquid, and on 31 October Sovryn was exploited and paused its whole protocol. During the liquidity shortage I paused new deposits and schedule creation to protect users, and set up a bot to monitor liquidity and notify users when their positions became withdrawable.

V2’s contracts are written and tested. It launches with LayerBank and idle mode, without Sovryn. DOC opens at launch. The USDRIF and USDT0 handlers will be audited and deployed with it, and open for deposits once the off-chain controls that keep their purchases within a published price standard are in, funded by this grant. V2 gets two independent audits before it touches mainnet. BitChill has been accepted in audit grant programs by two security firms: Zealynx Security (50% grant) and Kann Audits (full grant with a 250 USDC participation fee). Kann’s security review starts before this vote ends; Zealynx then audits the code with Kann’s fixes in. This grant covers the rest of Zealynx’s audit, the relaunch, the USDRIF/USDT0 capacity controls, and a year of operation, with the first 90 days measured in public.

Founder and team

Antonio Rodríguez-Ynyesto. Blockchain architect and Solidity/backend developer, about three years in the field. From March 2025 to June 2026 I was a blockchain developer at Umbrella Research, where I built an automated Aave liquidation engine from an architecture defined by its founders, Carlos Buendía (formerly ConsenSys and Avantgarde) and José Carlos Montero Gómez (OpenZeppelin). The project wound down after Chainlink SVR changed the economics of third-party liquidations.

GitHub username: Ynyesto · Forum: BitChill

BitChill started as the first-prize project of the Vottun & DexTools hackathon, sponsored by Rootstock, in January 2024. I put together its four-person team and became its tech lead; the other three have since left. I now build and operate everything alone: contracts, keeper, backend, frontend, infrastructure and user support. All infrastructure (domain, servers, email, deployment keys) is under my control and documented in a bootstrap repository, so a new operator could stand it up from scratch.

Trust model

  • Users’ funds: non-custodial. The contracts can only spend each user’s stablecoins on rBTC purchases at the amount and frequency the user set. Withdrawals of stablecoins, yield and rBTC are user-only.

  • Admin: each V2 contract has a single owner, a 2-of-3 Safe on Rootstock. I hold all three signer keys, each on its own hardware wallet, stored in separate places. Ownership moves in two steps (propose, then accept) and cannot be renounced. The owner can:

    • add a new route, meaning a place a stablecoin’s deposits can go (idle in the contract, or a lending venue such as LayerBank), and assign each stablecoin’s route its handler contract, once;
    • pause and unpause new deposits for one stablecoin on one route;
    • add and revoke keeper addresses;
    • set each handler’s fee parameters, with the maximum rate capped in code at 5%, and change the fee collector address;
    • set the minimum purchase amount per token, the minimum purchase period (whole days) and the maximum schedules per user and token;
    • on DEX handlers, allowlist swap paths, set the oracle-based slippage floor, and replace the BTC/USD oracle the floor is computed from.

    What admin cannot do: the contracts are not proxies, so there are no upgrades. The owner cannot replace or remove a handler once it is assigned. A handler is retired by registering a new route, and users then exit and re-enter by hand. No function lets the owner move, migrate or rescue user funds, and every payout goes to the schedule owner. There is no function to pause withdrawals or existing purchases; the owner’s only pause is on new deposits, one token and route at a time. In yield mode, stablecoin withdrawals depend on the lending venue’s liquidity and on the venue not being paused. If the venue can’t pay out stablecoins, users can instead withdraw their lending shares (the venue’s receipt tokens) to their own wallet and redeem them with the venue later; a paused venue may block that too.

  • Keeper: an off-chain job that runs once a week and executes due purchases in batches. On-chain it is an allowlisted address the owner can revoke at any time. For every purchase the contract reads the buyer, the amount and the due date from the user’s stored schedule, so the keeper cannot spend more than the periodic amount, buy early or credit another account. The whole batch reverts if any purchase fails.

    The keeper is trusted for liveness, not custody. If it skips a schedule, a purchase is delayed but nothing is lost. A leaked keeper key could also hold stablecoin withdrawals until revoked, activating back-to-back the short window that blocks schedule changes (inactive by default). The window is a safety mechanism so nobody can front-run batches to make them revert: nobody gains from doing that, but it would stall the protocol, which is why this guard was implemented. rBTC withdrawals stay open, and the key cannot move anyone’s funds. At launch DOC is converted through Money On Chain’s redemption at its oracle price, so there is no pool a leaked keeper key could trade users against. On DEX routes, once enabled, an oracle-based floor (3% below the oracle at launch settings) limits what a leaked key could cost per purchase until the owner revokes it. The floor is a setting the owner controls, not a fixed limit.

Product

Problem. Buying a fixed amount on a schedule takes the timing decision out of buying bitcoin. On Rootstock there is no simple, non-custodial way to do that, and stablecoins waiting for the next purchase sit idle.

How it works.

  1. The user connects a Rootstock-compatible wallet, picks an amount per purchase (minimum USD 25) and a frequency (every 1, 2 or 4 weeks).
  2. They deposit DOC and choose yield mode, where BitChill supplies it to a lending venue, or idle mode, where it stays in the contract.
  3. Each week the keeper withdraws only what due purchases need and buys rBTC for everyone in one transaction. DOC is converted through Money On Chain’s redemption (primary market), not a DEX.
  4. The user withdraws rBTC, remaining stablecoins and accrued yield at any time. In yield mode, stablecoin withdrawals depend on the lending venue’s liquidity.

Fees. V1 charges 1% per purchase. V2 charges 1% on purchases up to USD 250, then a rate that falls smoothly toward 0.2% as the purchase grows: 0.8% at USD 500, about 0.55% at USD 1,000 and about 0.24% at USD 10,000. Each purchase is priced on its own size, regardless of the deposit behind it or the user’s other schedules. The owner can change the curve, but the maximum rate is capped in code at 5%. V2 takes its fee in rBTC, as a share of what each purchase produces, so BitChill accumulates BTC on the same schedule as its users. BitChill pays the gas for scheduled purchases; users pay gas only for their own deposits, withdrawals and schedule changes.

What changes in V2.

  • LayerBank, not Sovryn. V1’s lending venues have failed (see Disclosures). V2 launches yield mode on LayerBank. New lending venues can be added as new routes by deploying new handler contracts for each venue + stablecoin combination.
  • Idle mode. A schedule can stay out of lending entirely.
  • Pause. Some V1 users withdrew everything to stop buying for a while, then deposited again later. V2 lets them pause a schedule instead, so their funds stay put and keep earning yield.
  • DOC at launch; USDRIF and USDT0 under a published price standard. DOC buys rBTC through Money On Chain’s redemption at the oracle price less a 0.15% fee (set by Money On Chain), with no price impact, as long as redemption is open. USDRIF and USDT0 go through Rootstock’s Uniswap pools. At block 9,406,075, against the oracle price and before BitChill’s fee, a small USDT0 swap got about 0.5% less rBTC (0.3% pool fee, 0.2% pool premium) and a small USDRIF swap about 1.05% less (0.35% fees, 0.2% premium, 0.5% USDRIF discount). Larger purchases add price impact, about 0.28% at USD 2,000. BitChill can’t change the premiums, and users will see them before depositing. What it controls is the size of its batches, without changing the audited contracts: the keeper and a data service admit a schedule only if every projected weekly purchase of its pool stays under 0.5% in pool fees plus impact, queue the rest, and reject any fill above 0.75%. BitChill’s own fee comes on top of both limits. Both coins buy rBTC through the same pool, and at the sampled block the 0.5% limit lets it take only about USD 1,000–1,500 of their combined purchases each week (the lower figure if it is all USDRIF, whose fees are higher). So USDRIF and USDT0 stay small on purpose. At the sampled block, small purchases meet the standard; the controls set each pool’s capacity and recheck the quality of every fill. The grant pays for this work, and the full policy is published with M1.
  • Hardening applied during the redeploy.
    • Price protection on swaps. Each USDRIF or USDT0 purchase has a minimum amount of rBTC it must return: the higher of what the keeper asks for and a floor derived from the Money On Chain oracle price. After the swap, the contract measures the rBTC that actually arrived and reverts the purchase if it is short. DOC purchases have no pool that could be moved: Money On Chain redeems at its oracle price, and the contract still checks the rBTC received against the keeper’s minimum.
    • Deposit and re-entrancy safety. A deposit that arrives short of the requested amount, as with a fee-on-transfer token, is rejected. Every user function that changes a schedule is re-entrancy guarded, and deposits and withdrawals update balances before any external call.

Integration with Rootstock. rBTC is the asset accumulated and the currency of BitChill’s fees. The inputs are Rootstock’s three main stablecoins: DOC, converted through Money On Chain’s primary market, and USDRIF and USDT0, through Rootstock’s Uniswap pools. Deposits in yield mode add supply to Rootstock lending markets.

Security

  • V1 audit: Ivan Fitro, independent security researcher, April 2025, with a follow-up review of the mitigations in June 2025 (reports), paid from the first tranche of the 2024 Rootstock grant.
  • First V2 audit: Kann Audits, through its audit grant program, which covers the full cost except a USD 250 participation fee. I pay that fee myself; it is not charged to this grant. Scope: the V2 contracts in src/, excluding src/tropykus-legacy/ and src/sovryn-legacy/, test-only adapters that no deploy script can put on a live network. It runs from 17 November 2026 for 10 days, before the M1 vote ends, and includes one round of fix review; the fixes go into the commit Zealynx audits. The report is published in the repository’s audits/ folder.
  • Second V2 audit: Zealynx. Scope: same as above. Commit: frozen at a public tag as part of M1. Price: USD 6,000 for one auditor-week (5 working days), for about 1,530 lines of implementation. Zealynx’s Growth grant covers 50%, USD 3,000, and is held for BitChill until 31 December 2026. This grant covers the remaining USD 3,000. The audit starts once M1 is funded; the date is fixed at signing, and the report follows a week later. Fix review is included, for as many rounds as it takes to fix the reported issues. The report is published unredacted, hosted by the DAO and on Zealynx’s Audits page.
  • Testing: about 850 unit, integration and fuzz tests, 11 stateful invariants (64 runs of depth 512) and a 106-property Medusa fuzzing suite. They run across every combination of purchase venue, lending protocol and stablecoin, and against forked Rootstock mainnet state (510 passing tests per lending venue). There are also 42 gas benchmarks and 16 mainnet-fork probes. Counts are at commit 6b2ed77 and will be updated at the audit tag.
  • Automated pre-audit scans: Zealynx’s Krait, Nethermind’s AuditAgent and Pashov Audit Group’s open-source auditor were run on the V2 code; reports and dispositions are public in the repository’s audits/ folder. They are not a substitute for the manual audit.
  • Monitoring: during the Sovryn liquidity crisis I built on-chain monitoring of its free liquidity, loan maturities, collateral mix and overdue loans. The LayerBank equivalent is part of this grant’s work and runs before V2 opens deposits.
  • Rule: production contracts are deployed only at the commit Zealynx reviewed, with all fixes re-reviewed. Kann reviews an earlier commit, and its fixes are in the code Zealynx audits. The deployment script and its parameters are not audited: the deployment test checks the wiring, and the parameters are published with M3’s evidence. The off-chain capacity controls are tested and reviewed, not audited.

Traction and demand validation (V3.3 Section 3.1)

BitChill is a consumer application, so the evidence is usage data, which V1 has a year of. TVL undersells a DCA protocol: deposits get spent on rBTC as schedules run, so what flowed through BitChill is far larger than what sat in it at any moment.

Metric V1, 10 Sep 2025 – 5 Oct 2026 (last purchase before the wind-down)
DOC spent on rBTC purchases ~USD 60,000
rBTC purchased for users ~0.8 rBTC
Scheduled purchases executed 1k+, in 90+ batches
Successful transactions to DcaManager 717: 617 from 73 user wallets, 99 from the keeper, 1 admin
Peak deposited principal (excluding interest) ~USD 26,000
Wallets/schedules with at least one purchase 56/103
Wallets/schedules with a schedule buying for 30+ days 37/58
  • Recurring by design: 37 of the 56 wallets that made a purchase (66%) had a schedule that kept buying for 30+ days, and 13 of those ran two or more such schedules.
  • Schedules that ran to the end: the longest made 52 weekly purchases of 25 DOC between 6 October 2025 and 5 October 2026. Nine schedules, from seven wallets, kept buying for 266 days or more. Three ran until their deposit was spent; the other six were withdrawn in early August 2026, when doubts about Sovryn spread.
  • Concentration: the top four wallets made 43% of the volume; the median wallet spent about USD 435.
  • TVL today. DefiLlama shows under USD 1,000. The drop came with the venue failures and the deposit pause, which removed the reason to keep funds in V1.

Purchases are executed by the keeper in batches, so they don’t show up as transactions from users’ wallets.

Getting users back. V1 users left because of the venues, not the product. One wrote to me in August (translated): “I’m really happy with BitChill, and I only left because of the latest Sovryn news. Keep me posted, I’d like to use it again.” At relaunch I’ll contact some V1 users directly and announce V2 on X. Manu Vega, marketing assistant at Money On Chain and Rootstock and RootstockCollective ambassador, has invited me to present BitChill in a Space with Rootstock’s ambassadors. He is also asking Money On Chain about an episode of Charlas con Sato, its community’s YouTube show. Milton Rodríguez, Rootstock Ambassador, has confirmed he will share the V2 launch. DOC holders are the natural first audience: V2 launches with DOC, the stablecoin they already hold, and every DOC purchase is a redemption through Money On Chain.

Cross-funding disclosure (V3.3 Section 6)

Funder When Amount Scope Status
Rootstock 2024 USD 15,000 total, split over 3 payments V1 build and audit Payment 1 (USD 5,000) received and spent on V1’s development and audit. Payments 2 and 3 not received due to missed deadlines.
Zealynx Growth Grant Oct 2026 USD 3,000 (50% of the V2 audit, as a discount) V2 audit Granted; held for BitChill until 31 Dec 2026
Kann Audits Grant Program Nov 2026 Full audit cost, except a USD 250 participation fee First V2 audit Accepted; fee paid by me, not by this grant

On the 2024 grant: BitChill was then a side project for all of us, and a large part of the work was learning how to build it. We missed the deadlines and forfeited two thirds of the grant, but still shipped V1, which I have run alone since late 2025. Two things are different now. BitChill is now my main project. And V2 is already written and tested: what this grant pays for is audit fixes, deployment and operation, which I have done before. This proposal is new funding for new scope, not a continuation of that grant.

Prior security incident disclosure (V3.3 Section 8)

No exploit of BitChill’s code and no loss of funds from a bug in it. Three events at the protocols BitChill depends on affected users and have to be on the record.

  1. Sovryn: illiquid DOC pool from late August 2026, then an exploit and a protocol-wide pause on 31 October. V1 supplies DOC to Sovryn’s iSUSD pool. From 27 August, when a single depositor withdrew 70,000 DOC within ten minutes, the pool was almost fully lent out, so lenders, BitChill included, could only withdraw what borrowers repaid. When a weekly batch lacked liquidity, the keeper retried as some freed up, and every remaining schedule still got its weekly purchase until V1’s wind-down, announced on 7 October. On 31 October Sovryn’s lending and margin contracts were exploited, and on 1 November its XUSD/WRBTC AMM. According to Sovryn’s 8 November disclosure, “Sovryn attack disclosure: BOB, Rootstock and Babelfish”, both came from the same kind of flaw: accepting a reported token amount without checking that the tokens arrived. The disclosure puts the lending outflow at about 670,775 XUSD and 4.03 WRBTC and the AMM loss at about 542,184 XUSD and 6.65 WRBTC, and says these are not final user losses. Lending, borrowing and margin remain paused, including withdrawals from the lending pools, with no reopening date (recovery manifest, “Sovryn mitigation and recovery manifest”, 9 November). On-chain, the DOC pool V1 uses (iSUSD) was not hit, but it is frozen by the pause. Separately, since 26 October Sovryn holds lending withdrawals for 24 hours (SIP-0096), and V1’s handler was not built for delayed withdrawals, so principal withdrawals revert even when the pool has liquidity.
    Impact: about 777 DOC remains in Sovryn: 655 DOC of principal across three users and 122 DOC of unwithdrawn interest. Two users have active schedules; the third has 2 DOC of principal remaining. Another 37 users have only unwithdrawn interest. Seven users have at least 1 DOC remaining in total; the other 33 have less than 1 DOC each, 21 of them below 1 cent. Recovery depends on Sovryn resuming withdrawals and available liquidity. rBTC already bought is not affected and can be withdrawn.
    Remediation: new deposits and schedules in V1 are paused. Once Sovryn unpauses, I will put a small DOC buffer in V1’s handler at my own cost so these users can withdraw in daily rounds that work around the 24-hour delay. During the illiquid period, a bot I set up notified users when liquidity freed up; it was available to every affected user on request, and one used it to withdraw everything. The two users with active schedules never requested access. V2 launches without Sovryn.
  2. Tropykus shutdown (announced 28 April 2026, support ended 31 August 2026). Tropykus announced its shutdown on 28 April. V1 also supported USDRIF and DOC through Tropykus. Users had withdrawn their funds well before Tropykus took its interface down, and since BitChill’s Tropykus positions were held through its own contracts, withdrawal stayed possible regardless. In fact, a user who had left ~2 DOC of accrued interest behind withdrew them on 4 November.
  3. Money On Chain pause (16–26 April 2026). On 16 April Money On Chain activated its Panic Button after a critical flaw in its oracle was reported, and paused DOC minting and redemption until a governance-approved fix was applied on 26 April (Money On Chain’s “Panic Button Activation — Incident Report”). No funds were lost and the flaw was not exploited. V1 buys rBTC by redeeming DOC, so the DOC purchases due on Monday 20 April were skipped; purchases resumed on 27 April. No funds were affected: the DOC stayed in users’ schedules, which simply ran one week longer. V2 depends on Money On Chain redemption in the same way.

Grant request and milestones

Total USD 20,000, paid in rBTC (see Payment rail), Strategic tier, one on-chain proposal per tranche. Initial tranche 20%, terminal tranche 30% withheld until the post-completion impact review.

Milestone Amount (USD) Releases on Deliverables and evidence
M1 4,000 (20%) Execution of the M1 on-chain vote V2 code frozen at a public tag; Zealynx engagement signed with a start date; audit scope document public; DEX capacity policy published
M2 5,500 Audit delivered and fixes reviewed Zealynx report published unredacted and hosted by the DAO; all critical and high findings fixed and re-reviewed; open medium/low findings listed with reasons
M3 4,500 V2 live on mainnet V2 deployed at the audited commit and verified on Blockscout; DOC live on LayerBank and in idle mode; USDRIF and USDT0 handlers deployed, deposits closed until their controls are in; withdrawal page for V1 users; public launch announcement on X and on this thread (day 0 is the first Monday after it); public script that reproduces the monthly figures from chain data
M4 6,000 (30%) Day 90 after relaunch, then the impact review Day-90 report against the targets below; USDRIF and USDT0 open for deposits under the price standard (below)

M4 is sized to day-90 volume, not all-or-nothing. Day 0 is the first Monday, BitChill’s weekly purchase day, after I announce the relaunch on X and on this thread, which opens V2 to everyone. Any test purchases I might make on mainnet before then don’t count. Below USD 3,500 of stablecoin volume converted to rBTC by day 90, M4 is not requested. At or above it, the amount is 0.6 × the volume in USD, rounded down and capped at USD 6,000: USD 2,100 at USD 3,500, the full USD 6,000 at USD 10,000 or more. Test activity is excluded. M4 also requires USDRIF and USDT0 to be open for deposits by day 90, with the capacity controls running and their acceptance test results published. The keeper and data service stay private, like the rest of BitChill’s off-chain code; the deposits, purchases and their prices are all on chain. The one exception: if either coin is still closed at day 90, M4 is the amount above minus USD 4,500, that work’s budget line, and never below zero. For example, USD 10,000 of volume would then give USD 1,500 instead of USD 6,000, and with USD 7,500 of volume or less there would be no M4 at all, since 0.6 × 7,500 is exactly the USD 4,500 deducted. If the controls are running but the published capacity probe shows a pool can’t take even a minimum purchase within the standard, the coin stays closed for market reasons and there is no reduction. The figures, the block and the arithmetic go into this thread before the M4 on-chain proposal is filed.

Use of funds. The work lines pay for my own time at my usual rate of USD 75/hour: 220 hours, USD 16,500. The work will take more hours than that, and I don’t charge for the extra ones. The audit and operations lines are passed through at cost.

Item Hours USD
V2 audit, my 50% share, passed to Zealynx at cost 3,000
Fixes and fix reviews for both audits 40 3,000
Deployment, LayerBank integration and monitoring, V1 withdrawal page 40 3,000
USDRIF/USDT0 capacity work: keeper, data service, front end, docs 60 4,500
Reporting script, Monthly Activity Reports, AMAs and user support 30 2,250
Maintenance over 12 months 50 3,750
Operations over 12 months: keeper gas, server, RPC, domain, X subscription 500
Total 220 20,000

Payment rail. Tranches in rBTC to the grant-only wallet submitted in the application, the same wallet for every tranche. I suggest each on-chain proposal converts its USD amount to rBTC at the Money On Chain BTC/USD oracle price at a block stated in the proposal, unless the RootstockCollective prefers another method. Either way, the DAO pays a fixed USD value and any price move before payout is mine. V3.3 makes USDRIF the default; I ask for rBTC because BitChill runs on it: keeper gas for every scheduled purchase is paid in rBTC, and BitChill’s fees are earned in it. The part that pays Zealynx, in USDC, I will swap and bridge to Base. Zealynx is paid USD 1,500 at signing and USD 1,500 when the report is delivered, before M2 can release. M1 covers both payments; I cover the bridging costs.

After the grant. In V1’s first year, fees covered BitChill’s running costs and a little more. Until V2’s fees do the same, I cover any gap myself.

Performance targets

Rootstock’s grants team has agreed that BitChill’s FES should take into account that its TVL falls by design, as deposits are spent on rBTC. I propose stablecoin volume converted to rBTC through BitChill as the headline, with transactions from user wallets and wallets with a schedule buying for 30+ days alongside. TVL is reported too, but isn’t the headline (see Traction). V1’s year shows the demand; it is not counted as impact of this grant.

Metric V1 baseline Day 90 after relaunch Month 12
Stablecoin volume converted to rBTC ~USD 15,000 a quarter on average, ~USD 60,000 in year one USD 10,000 USD 40,000
Transactions from user wallets to DcaManager 158 by day 90, 617 over V1’s life 105 400
Wallets with a schedule buying for 30+ days 16 by day 90, 37 over V1’s life 11 25

The targets are about two thirds of the V1 baselines in the table. V2 has two independent audits and a year of track record, but it restarts from zero deposits, on one lending venue, with lower yields than V1, so I am not assuming it beats V1. If volume counts dollar for dollar, two thirds of V1’s year, USD 40,000, is 2× the grant, the minimum FES for the Strategic tier, before counting retained TVL and new wallets. Keeper transactions and rBTC purchased are reported but not targeted: V2 buys for every route in one transaction a week where V1 needed one per route, and rBTC purchased moves with the BTC price.

Monthly Activity Reports go on this thread between milestones, with TVL delta, transaction count and new wallets.

Alignment with 2026 Strategic Themes

  • Rootstock onboarding UX. Set once, runs weekly with no further transactions from the user. It is one of the simplest first things a new user can do on Rootstock.
  • USDRIF adoption. The USDRIF handler is implemented and included in both planned V2 audits. It opens for deposits by day 90 at the latest, once its capacity controls are in (see M4).

Risks

  • Venue risk. Rootstock lending is thin, and its oldest venue has just been exploited and paused. For the foreseeable future, yield mode has one venue, LayerBank, whose admin (which can upgrade its pool proxies) is an EOA; I raised this with the LayerBank team through Rootstock Labs’ Head of Integrations, Alejandro Cavallero, and they chose to keep it. If a venue runs short of free liquidity, purchases and withdrawals on that route wait until it frees up, as happened with Sovryn in V1; idle schedules are unaffected. The owner can pause new deposits into a route that degrades, and the monitor watches its liquidity, but V2 cannot recover what a venue loses. Idle mode avoids that exposure entirely.
  • Low yields. If venue rates stay near zero, the product is still DCA into rBTC; the yield is a bonus, not the pitch.
  • Solo builder. The only part that depends on me day to day is the keeper. If it stopped, purchases would pause, but users could still withdraw, subject to the lending venue’s liquidity in yield mode. Tranches release only on delivery: if I stop, the DAO pays nothing further, and what it has paid has bought delivered work.
  • V1 funds still in Sovryn. Positions currently account for about 777 DOC: 655 DOC of remaining principal across three users (one of them with just 2 DOC) and 122 DOC of unwithdrawn interest. Seven users have combined principal and interest of at least 1 DOC remaining; another 33 have less than 1 DOC each, including 21 with less than one cent. Recovery depends on Sovryn resuming withdrawals and available liquidity (see Disclosures). A dedicated withdrawal page stays up for at least three months after V2 launches or Sovryn resumes withdrawals, whichever happens later, unless all remaining user funds are withdrawn sooner.

Timeline

Approximate, gated on votes and on the auditor’s slot.

  • Mid-November 2026: proposal posted, 1–3 weeks of discussion, Initial AMA.
  • 17–27 November 2026: Kann Audits review (10 days), then fixes and Kann’s fix review, before the Zealynx commit is frozen.
  • December 2026: 7-day on-chain vote for M1 and execution; Zealynx engagement signed and first payment made. The M1 vote must execute before 31 December, when Zealynx’s grant hold ends.
  • Late December 2026 – January 2027: audit (one week to the report), fixes and fix review (M2).
  • February–March 2027: V2 mainnet relaunch (M3).
  • Relaunch + 90 days, around June 2027: day-90 report and impact review (M4).

Obligations

Monthly Activity Reports until M4, milestone AMAs, the final retrospective AMA, and the post-grant obligations in V3.2: accepted. After M4 I keep posting monthly updates on this thread until 12 months after the grant closes, and the script that reproduces the figures stays public.

Links

Conclusion

V1 showed that people on Rootstock will DCA into rBTC: 56 wallets turned about USD 60,000 of DOC into ~0.8 rBTC in a year, until the lending venues underneath them failed. This grant audits and relaunches V2 without Sovryn, with an idle mode that carries no lending risk, and ties its final payment to usage and to opening USDRIF and USDT0.

Great to see that this project has everything ready for v2, Antonio! I think this is a very important step for the ecosystem.

I think the inactivity periods are a good idea, since many users don’t have the same confidence in current lending platforms that they had in Tropykus.

As you know, I’m a big fan of DOC, but adding USDRIF and UST0 also gives users more options. On that point, I have a quick question: in v1, BitChill would accumulate a certain amount of DOC before making a deposit or swapping it for RBTC to minimize the impact of network fees. In this version, liquidity will come in through three different tokens, so how will you handle that? Will you wait until there’s enough liquidity in each token to keep the impact of fees as low as possible?

The amount you’re requesting seems quite high to me, but it’s great that you’re getting an audit. I understand that this adds USD 3,000 to the costs. Since I’m not a technical person, just a regular user, I’d be interested in hearing what others think about the team conducting the audit.

If the DAO delegates trust the auditors, then my vote is obviously in favor. And of course, you can count on the Rootstock and Rootstock Collective ambassadors to help spread the word.

That’s all from me: just those two small questions, and congratulations on getting the project to this stage and sharing it here on the forum!

1 Like

Hi Antonio, thanks for such a detailed proposal. It’s rare to see incidents, a forfeited earlier grant and venue risks set out this openly, and it makes the proposal much easier to assess. A few points I don’t think have been raised yet:

1. The M4 payout formula can be met with your own money. The M4 payout is 0.6 × volume, up to USD 6,000. That means about USD 10,000 of volume releases the full tranche. Fees are paid to the protocol owner, so pushing your own stablecoins through BitChill would cost very little: some slippage and gas. I’m not suggesting that’s the plan, but the design should rule it out. Could you count only wallets with no link to the team or the Safe, cap how much any one wallet can add to the M4 volume, and publish a wallet-level breakdown with the day-90 report? Another option would be to require a minimum number of distinct wallets before the volume-based payout applies at all.

2. The 2-of-3 Safe works as a 1-of-1. You hold all three signer keys, so the multisig gives no protection against a compromised device or a single bad decision. I’d suggest handing at least one key, ideally two, to independent people before mainnet relaunch, and making that an M3 deliverable. It would also help with the solo-builder risk, because someone else could act if you were unavailable.

3. Fee changes need a delay. A 5% hard cap is good, but the owner can still change the fee curve at any time. A timelock, or a published notice period (say 14 days) before any increase, would let users with live schedules pause or withdraw first. For a DCA product, where people set up a schedule and leave it running, that matters more than usual.

4. The order of the two audits. Kann runs from 17 to 27 November, and the M1 code freeze comes after that. Could you confirm that the frozen tag will include all Kann fixes, and that Zealynx will audit exactly that commit? It would also help to say how any change made after the Zealynx review will be handled, perhaps by limiting fixes to items Zealynx has reviewed. One auditor-week for about 1,530 lines is tight, so I’d like to know whether the fix review is a real second pass or a quick check.

5. Default to idle mode. V1’s TVL was hit by lending venues failing, and LayerBank’s pools can be upgraded by a single key. Given both, I’d make idle mode the default, require an explicit opt-in for yield with a clear risk warning in the interface, and publish a short plan for what happens if your LayerBank monitoring sees something wrong. Users can’t count on the owner to move their funds, so they need to know quickly when to withdraw.

6. No funding is tied to months 4–12. The budget covers 12 months of maintenance and operations, but the last payment comes at day 90 and the month-12 KPIs aren’t tied to any funding. Would you hold back a small part of M4 (say USD 1,000) until a month-12 report, or at least commit to monthly reports for the full year? I’d also add a retention measure, such as the number of schedules still running at month 12, alongside volume.

7. V1 funds stuck in Sovryn. The withdrawal page stays up for three months after relaunch, but Sovryn may be paused for longer. Could you commit to keeping it live until those three users can actually get their DOC back?

8. Lessons from the 2024 grant. You’ve said the deadlines were missed. A line or two on what is different this time would reassure voters. The milestone structure already helps, but a direct answer would be better.

Overall, this looks like a careful relaunch of a product with real usage behind it. I’d be happy to support it, especially if points 1 and 2 are dealt with.

1 Like

Hi, Manu! Thank you for your insights and questions!

Just to clarify, BitChill didn’t wait for DOC to accumulate before depositing it or executing purchases. Batch purchases were executed as they were due, even when that meant (as it did at the very beginning and in the last couple of months, when purchase volume was way below its peak) that the fees earned by BitChill were lower than the gas cost.

You are right that the purpose of batching purchases is exactly to lower the gas cost per individual purchase, and you are also right to question how this works with three tokens. In V2, we have a new function that batches batches. A “metabatch” function, if you will. That way, due purchases from schedules using any of the three stablecoins, in either mode (idle or LayerBank), can all be executed in a single transaction.

Of course, there’s a limit, since the transaction has to fit within the block’s gas limit, but V1 always stayed well below that limit. If BitChill ever gains enough traction for the gas limit to become relevant, we can split the purchases across as many transactions as necessary. USDRIF and USDT0 purchases are also capped to protect their prices, given the liquidity in their pools, as described in the proposal.

Hi, Seviramar! Thanks a lot for such a detailed response! I’ll address every point below.

1. M4 volume

That’s a fair point. The intention is to measure actual user activity, so my own purchases should not help unlock that payment.

I’ll explicitly exclude purchases funded by me, the admin Safe, the grant wallet, or any other wallet I control, as well as test activity. The day-90 report will include a wallet-level breakdown showing the volume counted, the exclusions, and the resulting M4 calculation. I’ll publish that here before submitting the payment proposal, so delegates can check it.

I would prefer to start with those exclusions and transparent reporting rather than introduce a per-wallet cap. A cap could discount legitimate users while still being easy to bypass with multiple wallets. I also cannot prove that every apparently unrelated wallet belongs to a different person.

2. The Safe

I understand your point, but I think keeping the current setup is reasonable in this case. BitChill was designed as a trust-minimized protocol: the protection comes primarily from what the contracts allow, rather than from trusting whoever holds the admin keys.

The contracts are not behind proxies. The admin cannot upgrade them, replace an assigned handler, withdraw users’ funds, or move their funds at all. Users decide whether to enter a new route themselves.

The Safe handles limited administrative functions, such as rotating the keeper, pausing new deposits, registering new handlers, and configuring fees and price protections. Those powers exist, but they are very different from having control over users’ balances or the ability to rewrite the protocol, and the fees are capped at 5% (and aren’t meant to ever be set anywhere near that level).

I expect very little use for the Safe. For example, if another lending protocol launches on Rootstock, I could deploy an integration and use the Safe to register it. Requiring another person’s signature for that operation seems like unnecessary friction given that existing users’ funds would not move.

Holding the three keys on separate hardware wallets protects against a single compromised signing device. It does not provide independent oversight, but that is not the purpose of this setup. If the admin could upgrade contracts or directly control users’ funds, I would agree with your recommendation. Given BitChill’s intentionally limited admin powers, I think the current arrangement is a reasonable compromise.

Every Safe transaction is public on chain, and I can also list each one in the monthly reports with a brief explanation if that helps.

3. Fee changes

I don’t intend to change the fee curve. Quite a lot of work went into finding one that I think is fair across different purchase sizes. Still, I understand why users would want advance notice if that ever changes.

I’m happy to commit to announcing any fee increase here and on X at least 14 days before applying it. I would include the new parameters and examples of their effect on different purchase sizes.

I’d rather keep this as a public operating commitment than add a timelock to the contracts, which are about to enter their first audit on a fixed commit.

4. The audits

Yes, the frozen tag will include Kann’s fixes and their fix review. Zealynx will audit that exact commit, and I’ll publish the tag and commit hash.

Zealynx’s engagement includes as many fix-review rounds as necessary to address the reported findings. I would not describe each round as another full audit: it is a review of the fixes and their implications.

The deployed version will include those reviewed fixes. If another change becomes necessary after the review, I’ll send it back for review before deployment rather than treat the earlier report as covering code the auditors have not seen. The final deployment commit and its relationship to the audited version will be public.

5. Yield mode

I want to keep yield mode as the default because earning interest between purchases is part of BitChill’s value proposition. The interface lets users choose LayerBank or idle mode through a dropdown, and the website explains how both work and their risks.

BitChill does not borrow against users’ deposits or take leveraged positions. Yield mode does expose deposits to LayerBank’s lending and smart contract risks, which is precisely why V2 also offers idle mode. Users who prefer to avoid that exposure can choose it.

I don’t like the EOA upgrade authority either. I raised it with the LayerBank team through Rootstock Labs, as disclosed in the proposal. I still think offering LayerBank alongside an idle option is a reasonable choice.

If monitoring identifies a material problem, such as insufficient withdrawal liquidity, a pause, or a suspicious upgrade, I’ll investigate and publish an update on X. Where appropriate, I can also pause new deposits into the affected route. Existing users retain control of their positions and decide whether to withdraw.

6. Months 4–12

The proposal already commits me to monthly updates after M4 until 12 months after the grant closes. So reporting does continue beyond day 90.

I’m happy to make the month-12 report explicit and include the number of schedules still active then, alongside volume and the existing measure of wallets buying for at least 30 days. Those measure different things, and both are useful.

I would prefer to keep the proposed payment schedule. M4 already depends on measured usage and delivery of the USDRIF/USDT0 work, while the maintenance and reporting commitments continue after payment. I also continued building and operating V1 after forfeiting the remaining payments from the earlier grant. I hope that track record gives delegates some confidence in the commitment.

7. The V1 withdrawal page

Definitely. I’ve updated the proposal: the withdrawal page will stay available for at least three months after V2 launches or Sovryn resumes withdrawals, whichever happens later, unless all remaining user funds are withdrawn sooner. If I eventually take it down, I’ll remain available on X and BitChill’s contact email to help anyone who still needs to withdraw.

When Sovryn liquidity was low, I set up a Telegram bot, available to users on request, to notify them when liquidity became available for withdrawals. I intend to maintain that level of support in V2, if not improve it.
I also checked the remaining positions while updating this section. The approximately 780 DOC included unwithdrawn interest belonging to users who had already withdrawn their principal. I’ve corrected the breakdown and user count in both relevant sections of the proposal.

8. Lessons from the earlier grant

The biggest differences are my experience, the amount of work already completed, and the time I can now dedicate to BitChill.

In 2024, we were learning how to build much of the product as we built it. It was a side project, and the milestones were set by the grants team. We accepted them, so meeting them was still our responsibility. We missed the deadlines and forfeited the remaining payments, but continued the work and shipped V1.

This time, BitChill is my main project, V2’s contracts are already written and tested, and I’m proposing milestones based on the work that remains. I also have a year of operating the product behind me. I’m confident about delivering the work within my control, while the proposal makes the dependencies on votes, auditor availability, and user adoption explicit.

Again, thanks a lot for your thoughtful comments and questions. I tried to make the proposal as transparent as possible, so I appreciate you raising these points. I hope these explanations, together with my track record with V1, make delegates comfortable supporting the proposal :slight_smile:

Thanks for the clear answer, Antonio! The metabatch makes a lot of sense, and it’s good to know purchases always went out on time even when the gas cost more than the fees. Looking forward to seeing V2 on mainnet, count on us to help spread the word :orange_heart:

1 Like

Thanks, Manu! Really appreciate it :folded_hands: