The Stellar Community Fund is excited to announce the results of SCF #45. This round, 40 projects were awarded across the Open, Integration, and RFP tracks, totaling approximately $4.04M worth of XLM*. From confidential-finance infrastructure to real-world-asset tokenization, cross-border payment rails, and agentic-payment tooling, this cohort reflects the growing depth and ambition of the Stellar builder community.
Here’s the breakdown by track:
- Open Track: 8 projects, $838,100 worth of XLM*
- Integration Track: 28 projects, $2,760,516 worth of XLM*
- RFP Track (Developer Tooling): 4 projects, $445,000 worth of XLM*
Open Track
The Open Track funds new products and protocols building on Stellar, reviewed by the delegate panel and the NQG community vote. 8 projects were awarded this round.
Cara7 — $127,900 worth of XLM*
Institutional infrastructure for tokenizing auto-loan receivableson Stellar, with a configurable Soroban payment-waterfall contract, EURC distributions via the Stellar Asset Contract, MPC custody and fiat ramp integration, and issuer/investor dashboards.
ZKELLA Protocol — $127,800 worth of XLM*
A confidential-finance layer for Soroban: a note-based token standard wrapping SEP-41 assets, ZK circuits with on-chain verifiers, shielded swaps, opt-in viewing keys for compliance disclosure, a note indexer, and a TypeScript SDK.
Pods — $125,000 worth of XLM*
A request-for-quote settlement primitive on Soroban where allowlisted market makers compete to fill signed buyer requests that settle atomically on-chain, enabling distribution of tokenized real-world assets on Stellar with built-in compliance controls.
Astrolabe — $117,900 worth of XLM*
A Stellar block explorer with a self-hosted data layer (Galexie, ClickHouse, Postgres) offering full-history backfill, contract event search, DeFi and holder analytics, alerts, SEP-55/58 contract verification, read-call forms, and a public API.
Principal — $116,000 worth of XLM*
A Soroban-native protocol that separates tokenized RWAs into principal and yield components, while inheriting each underlying asset’s native compliance controls (authorization and clawback).
Pollar — $89,500 worth of XLM*
An SDK giving Stellar apps embedded wallets and USDC charging via Google or email login, plus a unified fiat on/off-ramp API across seven anchors in six Latin American countries, with webhooks, reconciliation, and KYC tiers.
Strata — $69,000 worth of XLM*
A white-label Soroban vault kit for real-world-asset projects: asynchronous tokenized-vault contracts, a guarded valuation oracle, notice-period redemptions, allowlist compliance, and an LP dApp that teams configure and deploy under Apache-2.0.
Fermah Pay — $65,000 worth of XLM*
Seller-side agentic-payments infrastructure on Soroban: a prepaid multi-tenant billing ledger contract, time-bounded recurring USDC charges via the Stellar Asset Contract, a batching settlement daemon with an x402-compatible interface, and an open-source SDK.
Integration Track
The Integration Track funds teams integrating existing products and building blocks into the Stellar ecosystem. 28 projects were awarded this round.
BWB — $150,000 worth of XLM*
A Brazilian real-estate tokenizationplatform using Stellar as its rail, with Privy embedded wallets, Soroban offering and eligibility contracts, Circle CCTP Base-Stellar liquidity, BRL settlement via Avenia, and DeFindex integrations.
Sorted — $150,000 worth of XLM*
A stablecoin banking app on Stellar for users on Android Go, older, or basic phones with intermittent connectivity in Africa and South Asia, providing digital-dollar access without crypto-native wallet, gas, or bridging complexity.
Neko — $126,710 worth of XLM*
A non-custodial consumer web app giving Latin American users access to real-world-asset and stablecoin protocols on Stellar. by composing audited building blocks (DeFindex, xBull Swap Kit) via their SDKs.
Balance — $125,000 worth of XLM*
Extending Balance’s production institutional custody for XLM and USDC by integrating the Stellar Anchor Platform and Stellar Disbursement Platform, adding anchor and SDP adapters to its existing custody API.
Tilt Pay — $124,000 worth of XLM*
A live mobile payment app migrating its settlement infrastructure to Stellar, moving existing user flows to native USDC settlement with DeFi integrations and in-app swaps.
Urbanflip — $120,000 worth of XLM*
Tokenizing Madrid real-estate projects on Stellar with per-project subscription contracts and compliance-integrated onboarding.
Lunar Finance — $118,000 worth of XLM*
A multichain aggregation protocol integrating Stellar into its unified routing system, aggregating Stellar DEXs such as Aquarius, Soroswap, and StellarBroker to enable same-chain and cross-chain transaction routing.
Micro BE — $115,000 worth of XLM*
Moving marina utility payments on-chain on Stellar, replacing prepaid RFID-card water and electricity credits at connected bollards (925+ across French marinas) with on-chain settlement and traceability.
Janus — $110,376 worth of XLM*
Payment and settlement infrastructure for freight forwarders, bringing cross-border carrier payments and customer invoicing onto Stellar to replace slow, costly correspondent-banking rails.
Agama — $108,200 worth of XLM*
Infrastructure for tokenized real-world assets on Soroban, focused on private credit, with vaults, a synthetic-dollar module, and USDC entry via Stellar wallets, Circle CCTP, or a MoneyGram SEP-24 anchor.
Crebit Rate Locks — $100,000 worth of XLM*
An API-first layer for short-duration, USDC-settled FX rate locks, letting payment providers and market makers manage exchange-rate exposure with on-chain settlement..
Minisend — $100,000 worth of XLM*
A Stellar anchor and settlement rail letting users send USDC and receive local currency via M-Pesa, Airtel Money, or bank in Africa, using SEP standards and Circle CCTP liquidity.
Sendana — $100,000 worth of XLM*
A borderless banking platform built on Stellar using USDC rails for freelancers and remote workers across the Global South (starting in the Caribbean), adding bank-transfer offramps via Visa Direct push-to-card and business banking.
Liqvid — $97,500 worth of XLM*
A Stellar-native lending market letting verified investors borrow USDC against Liqvid-issued tokenized RWAs as collateral, with isolated markets on Alula, a contract-derived SEP-40 NAV oracle, compliance-gated liquidation, and cross-chain USDC supply via CCTP.
Catlog — $97,000 worth of XLM*
Integrating Stellar as the core settlement and payout layer of its payment infrastructure with Stellar USDC as the settlement asset, routing multi-chain stablecoin checkout payments into Stellar via NEAR Intents and Circle CCTP.
Haven — $97,000 worth of XLM*
A privacy-first crypto neobank running natively on Stellar, built on Soroban smart-account wallets with passkey (secp256r1) auth and on-chain recovery, integrating six SCF building blocks for everyday banking.
Kutana — $97,000 worth of XLM*
A B2B cross-border trade platform for African SMEs moving its milestone-based escrow with embedded coverage mechanism, holding escrow in USDC on Stellar via Trustless Work’s audited contracts with milestone-based releases.
PigFi — $96,060 worth of XLM*
A mobile savings app expanding its DeFindex integration into a configuration-driven multi-vault experience, adding a reusable backend vault registry and adapter, normalized transaction-state handling, and mobile vault discovery screens.
Nila — $95,000 worth of XLM*
A stablecoin lending facility for smallholder agricultural credit in Tamil Nadu, India, porting its member cash ramp (Anchor Platform), member wallets (Privy), and USDC capital path (CCTP) onto Stellar, with remote-sensing crop verification replacing the field agent.
lomi. — $95,000 worth of XLM*
A Stellar bridge and anchor for XOF/USDC for lomi., a live francophone West African (UEMOA) payment processor, connecting its mobile-money and card checkout to Stellar-based settlement.
lend.xyz — $92,900 worth of XLM*
A protocol bringing tokenized real-estate bonds on-chain via French-law SPVs with structured documentation, enabling users to participate in structured real-estate operations on-chain, with periodic distributions.
Onboard — $89,000 worth of XLM*
A cross-border payments platform integrating Stellar as a settlement network for its consumer app, business product, and API, paying out local currency alongside its existing EVM, Tron, and Solana rails.
Lendoor — $85,000 worth of XLM*
Uncollateralized on-chain credit on Soroban settling in USDC, with cross-chain USDC liquidity via Circle CCTP, gasless onboarding via Privy, DeFindex treasury routing, and a portable on-chain credit score built from repayment history.
Enable — $76,000 worth of XLM*
Migrating a group-payments product to Stellar, building a Soroban group-wallet contract for shared, non-custodial pooled-USDC accounts with m-of-n approval and contribution/disbursement logic.
LendWise — $66,670 worth of XLM*
A DeFi lending aggregator standardizing lending-rate data and centralizing cross-protocol position monitoring, with optimization and execution tools, integrating Stellar lending markets into a unified interface.
SwiftEx — $54,500 worth of XLM*
A live non-custodial multi-chain wallet and trading app adding native USDC onboarding via CCTP, best-execution routing across the SDEX, Aquarius, and Soroswap, and DeFindex integrations, unifying trading, DeFi access, and P&L tracking on Stellar.
BIM Exchange — $39,600 worth of XLM*
A live multi-chain DeFi aggregator expanding into a multi-source Stellar execution layer, adding Privy embedded onboarding, Aquarius and SushiSwap liquidity, MoneyGram cash ramps, and a best-execution routing engine across Stellar venues.
Slimepay — $35,000 worth of XLM*
Slimepay’s Stellar anchor for certain African currencies, using the open-source Stellar Anchor Platform (SEP-24/SEP-6) so users deposit USDC or XLM and off-ramp to local bank rails.
RFP Track (Developer Tooling)
Targeted developer tooling needs, reviewed by the delegate panel only. Every RFP Track submission responds to a specific open RFP published for the quarter. 4 projects were awarded this round across two RFPs.
x402 Facilitator with Bazaar (discovery) support
An x402 payment facilitator plus a Bazaar discovery layer for Stellar, so AI agents can discover, price, and pay for services per request.
Rail402 — $125,000 worth of XLM*
An Apache-2.0 suite for discovering, purchasing, settling, and inspecting x402 services on Stellar, built on the @x402/stellar flow, providing standard supported/verify/settle facilitator endpoints and multiple settlement schemes.
AgentSmith x402 — $122,000 worth of XLM*
An Apache-2.0 x402 payment facilitator and Bazaar discovery layer for Stellar (testnet and pubnet) with non-custodial verify/settle and sponsored fees, letting AI agents discover and pay for services per request.
Rumble Fish — $78,000 worth of XLM*
An open-source x402 payment facilitator and Bazaar discovery layer for Stellar, letting AI agents find, price, and pay for API endpoints and MCP tools per request in USDC or any SEP-41 asset, with a metered “upto” settlement scheme and SDK support.
Stellar-compatible LayerZero DVN
A production-grade LayerZero Decentralized Verifier Network (DVN) for Stellar, giving cross-chain apps a second independent verifier on Stellar pathways.
Canary Protocol — $120,000 worth of XLM*
Extending Canary’s production LayerZero DVN to Stellar, giving cross-chain apps a second independent verifier with multi-VM message decoding (EVM, SVM, Move), Stellar-native signing, and TEE-isolated execution.
Referral Insights
The SCF referral program remains a powerful top-of-funnel for high-quality teams. Of the 40 projects awarded in SCF #45, 25 (63%) came into the program through a community referral rather than applying cold. Referred teams performed especially well in the Open Track, where 7 of the 8 awarded projects were referred.
A huge congratulations to all of these incredible teams! We’re excited to support you in driving meaningful, lasting impact on Stellar.
* The USD valuation of the Award in XLM is calculated using the CF Stellar Lumens-Dollar Settlement Price on the day of payment as administered, maintained, and reported by the cryptocurrency index provider CF Benchmarks Ltd. (using the ticker XLMUSD_RR) (available at https://www.cfbenchmarks.com/indices/XLMUSD_RR).
DISCLAIMER
** Project descriptions are based on information provided by award recipients or publicly available sources. SDF/SCF makes no representations or warranties regarding the accuracy, completeness, or compliance of any project description. Inclusion in this recap does not constitute an endorsement, recommendation, or investment advice by SDF or SCF with respect to any project, token, protocol, or digital asset. Readers should conduct independent due diligence.
Round Insights & Voting Dynamics
Since October 2023, the Stellar Community Fund (SCF) has utilized Neural Quorum Governance (NQG), a reputation-based voting mechanism to signal community sentiment in Award Allocation in the Community Vote phase of SCF Build (Open Track). Voting results are non-determinative, and final allocation decisions are made by the quarterly delegates panel.
Round at a Glance
SCF #45 broke every participation record: 172 voters from a 754-member community (22.81% turnout, second-highest ever) with 82.81% retention from #44. Voter count grew 34% round-over-round against 13% growth in the verified community, so participation outpaced recruitment rather than simply tracking it.
The result this round: 8 Open Track projects were awarded by the delegate panel. Turnout crossed 20% for the first time, and it did so while the ballot nearly doubled in size, with the community vote covering 44 Open Track submissions — the largest ballot since SCF #40. See all public data here and in the SCF Dashboard Analytics page.
The Community Signal
For the fifth consecutive round, every project the delegate panel awarded in the Open Track was also a top-ranked project in the NQG community vote — extending the 100% alignment***, a streak that began in SCF #41. All 8 panel awards were exactly NQG’s top 8 by community score.
While in the past few rounds the delegate panel has taken the >82% approved NQG as a guide to funding, the delegate panel approved two additional projects slightly below that percentage due to additional discussion and given the wide margin of the lowest-scoring awarded project (Principal, +615) sitting 95 NQG points above the highest non-awarded project.
Decentralization Measure
Total voting power grew from 589 to 1,075 (+83%) both due to overall voter growth and changes to the Trust Graph neuron (see detail in Neuron Performance section), and the distribution widened alongside it. The top 5 voters held 18.3% of all voting power, up from 12.1% in SCF #44; the top 10 held 32.2%, up from 22.3%. The Gini coefficient rose from 0.38 to 0.53 and the Theil index from 0.24 to 0.53.
The stratification is structural. Inequality within each tier is modest: Gini 0.30 among Pathfinders and 0.34 among Pilots. Almost all of the measured concentration is now between tiers: 33 Pilots (19.2% of voters) hold 56.9% of the voting power, while 131 Pathfinders (76.2% of voters) hold 36.9%. The average Pilot carries 6.1× the voting power of the average Pathfinder, up from 3.5× in #44.
This is the mechanism behaving as designed: Trust Graph is meant to reward accumulated peer trust, and peer trust accrues to long-tenured contributors. This does result in the headline inequality metrics to be at their highest recorded level, and mainly concentrated in active Pilots, which reflect the current state of the community. Additional work on the onboarding for Pathfinders to gain voting power and become Navigators (and eventually Pilots) is necessary to address voting power concentration rather than fixing parameter changes (see more detail in Looking Ahead section).
Quorum Delegation
NQG allows voters to delegate their vote to a quorum of trusted community members rather than voting directly on every submission. Each voter selects a panel of delegates (min. 7); if enough delegates participate and a supermajority agrees (>66.7%), the delegation resolves to Yes or No.
In SCF #45, delegation reached a new high: 87.6% of all vote-slots were delegated, up from 79.7% in #44. 139 of 172 voters (81%) relied entirely on delegation, 17 (10%) used a mix, and just 16 (9%) voted directly on every submission. With a 44-submission ballot, this is the expected and intended response: delegation is what lets a voter participate meaningfully across a doubled ballot without evaluating all 44 projects personally.
The system again performed flawlessly: all 6,630 delegated vote-slots resolved, a 100% resolution rate for the fifth consecutive round, with zero abstains in the final tally.
Important to note is that the active delegate pool contracted again, from 16 to 14, while the voter base grew 34%. Each delegate now carries roughly 474 delegated vote-slots on average, up from ~160 in #44 — a threefold increase in one round. The average quorum size rose slightly to 8.35 delegates (from 7.69), but with only 14 delegates available, quorums necessarily overlap heavily: the most-selected delegate appears in 145 of 170 quorums (85%), and the top five each appear in more than 70%. A handful of delegates going inactive in a future round would affect most quorums simultaneously, and it’s important to grow the active delegate pool.
Neuron Performance
NQG computes each voter’s influence through four “neurons” — modular reputation inputs that reflect different dimensions of a voter’s standing in the ecosystem (see exact parameters in the SCF Handbook).
The Trust Graph neuron was re-parameterised following the recommendation from the last round report to balance the Trust Neuron with other neurons following a UX update. The UX was updated to check Discord interactions with the user and suggest whom to add trust to, and invited all to update their trusted list of users prior to the Community Vote of #46.
After the #46 tally, the trust neurons were again adjusted to minimize extreme outliers caused by the highly-trusted bonus compounding without limit, which was not aligned with the intended behavior of the change. Every highly trusted voter (top 10% by PageRank) who trusted you multiplied your score by 1.15, once per truster, so a voter trusted by k highly-trusted peers earned 1.15^k. In this pre-publication tally run, one account trusted by 23 such peers reached a trust score of 82 against a PageRank ceiling of 3, taking 93.7 total voting power and 36% of all Trust Graph power. In this tally the Trust Loss Neuron was triggered for a number of users because of trust list changes of this powerful voter, which highlighted the issue: the system only took the amount of NQG ‘untrusted’ into account, not the amount of users who removed someone from their trusted list of users.
Parameter changes following initial tally:
- The bonus gain is now capped. The PageRank base is untouched; only the gain from stacked bonuses passes through a logistic curve, 20 × tanh(gain / 20). It is identity-like for small gains and saturates at 20, so compounding can no longer run away.
- The PageRank base was rescaled from 0–3 to 0–10, to give better separation between lightly and heavily trusted users now that the bonus no longer supplies that separation.
- Trust Loss now requires at least three distinct revokers, in addition to the existing NQG-sum threshold, preventing a single sufficiently powerful voter from triggering the full penalty alone.
All three are visible in the published data. The maximum trust score is ~33 — exactly the new theoretical ceiling of (10 base + 20 saturated gain) × 1.10 filled-trust-list bonus — and only one voter reaches it. Eleven voters score above 10, meaning their saturated bonus is doing real work without dominating. The top account’s share of Trust Graph power fell from 36.2% to 10.4%, and the top five from 73.7% to 44.0%.
Trust Loss applied to nobody this round: no voter had three or more peers revoke trust, so the neuron contributed exactly zero, and the minimum voting power in the round is 1.00 rather than the near-zero values the uncapped configuration produced.
The net effect on outcomes was small but real. The awarded set was unchanged — the panel’s 8 awards are NQG’s top 8 either way — but the ordering inside the top 8 shifted: Pollar rose from 6th to 4th, while Strata and Astrolabe each moved down one place. The gap between the last awarded and first non-awarded project widened from 66 to 95 points.
Neurons are now more balanced.
In SCF #45, 172 voters participated across three community tiers: roughly 76% Pathfinders (131), 5% Navigators (8), and 19% Pilots (33). Pathfinder growth again drove the round (88 → 131), while Navigator (7 → 8) and Pilot (33 → 33) counts were flat — the entire increase in the voter base came from the entry tier.
- Trust Graph — Now 48.6% of the average Pilot’s voting power, up from 10% in #44, and by far the neuron that separates Pilots from everyone else. Pathfinders draw almost none of their power from it (0.7%), which is expected: PageRank centrality and highly-trusted endorsements accrue over time. Even with the bonus capped, this remains the single largest structural differentiator in the system.
- Assigned Reputation — Still the base layer for lower tiers (58% of a Pathfinder’s power, 36% for Navigators) but only 23% for Pilots, down from 47%. The formula did not change; the denominator did.
- Vote Quality — Held its range in a seventh active round (max 4.77) with no negative scores for a second consecutive round. Coverage broadened notably: 122 of 172 voters (71%) now carry a non-zero value, up from 100 (58%) in the earlier data export, as more historical tranches were resolved.
- Voting History — Stable at 11.3% of round-wide VP on an unchanged ceiling. Pilots average 2.07 against 0.34 for Pathfinders, a ~6× gap that has narrowed from #44’s ~13× as more of the new cohort accumulated a second and third round of participation.
Across all voters, the Trust Graph rose in overall NQG from 8.8% to 29.6%, which now provides a meaningful balance with other neurons. It’s now the highest voting power contributor for Pilots, which as mentioned in the previous section, contributes to the increasing concentration of voting power. With increasingly intuitive UX and encouraging people to update their trusted lists more frequently, more Pathfinders and Navigators can be added to trusted lists.
The radar chart now highlights an interesting progression of what drives voting power: to start, Pathfinders build up voting power through voting history, followed by the outcome of their actions through voting quality, followed by gaining reputation in the community through tiers, followed by gaining trust from its existing community, the latter being the most difficult of all.
Looking Ahead
SCF #45 sets four new highs — the largest voter cohort in NQG history (172), the first round above 20% turnout, the strongest delegation share ever (87.6% of vote-slots), and the largest community-vote ballot since #40 (44 submissions). It also marks the fifth consecutive round of both 100% panel-NQG alignment and 100% delegation resolution, a stability streak now spanning well over a year of tallies.
It is also the round in which NQG’s trust layer was tested, found to have an unbounded failure mode, and repaired before publication. The compounding bonus and the single-revoker penalty were both caught in pre-publication analysis of the tally, fixed in the contracts, and the round was re-run. The mechanism worked the way a governance system should: the anomaly was visible in the data, and the fix shipped before the result was final.
Three questions carry into SCF #46:
- Tier stratification. The outlier problem is solved, but measured concentration is at a record high and is now structural: 19% of voters hold 57% of the power. A solution can be to more actively support Pathfinders in gaining voting power and becoming Navigators (and eventually Pilots), which requires clearer opportunities for Navigators, more visibility on the process, and increased interaction through improved UX.
- Delegate pool capacity. 14 delegates now serve 172 voters and carry ~474 vote-slots each, with the top delegate appearing in 85% of quorums. The delegate size has shrunk in two consecutive rounds (17 → 16 → 14) while the larger delegator pool has grown 72%. Recruiting and qualifying additional delegates for Q4 is the clearest structural risk to the streak.
- Signal versus capacity. With 14 positively-scored projects and 8 awards, and the exception made to deviate from the previously held >82% approved NQG score cut-off, the gap between community endorsement and funding changed in #46. Communicating what a positive NQG score does and does not promise will matter more as ballots change or grow over time.
For a full deep dive into NQG’s evolution, see the Two Years of Neural Quorum Governance retrospective. To learn more about how NQG works and how to participate, visit the SCF Handbook. If you’re interested in participating in the community vote and helping to shape the future of the Stellar ecosystem, become verified today and participate in the discussion on the Stellar Dev Discord!
SCF #45 Round Recap was originally published in Stellar Community on Medium, where people are continuing the conversation by highlighting and responding to this story.