Issue #326, backlog
#273 (docs/backlog.md).
Status: completed. The map is short, and several angles closed negative or
inconclusive — those are reported as such rather than padded. Every Wayfare
capability named below was checked against the tree at 295085b on
2026-09-25; every external claim carries its own source and the date it
was checked. No code was written, and none is proposed by this issue.
A numbering note, checked directly against the tracker rather than assumed. Issue #326's body says the map is "the input #157 needs to be worth answering." That reference is to backlog entry
#157— "Spike: who would consume this API, and what shape do they need" (docs/backlog.md, checked 2026-09-25) — which was filed as GitHub issue #217. GitHub issue #157 is a different issue entirely: "Depth at the dust size is the structural-floor probe", a V2 pricing issue, closed on 2026-08-31 (checked 2026-09-25 via the GitHub API). This is the same backlog-versus-GitHub numbering split documented in docs/sep38-african-fiat-survey.md. Issue #217 was closed on 2026-09-24 with an inconclusive "who" finding (docs/spike-api-consumers.md), which is exactly the gap this map supplies input to. The mismatch does not change this document's task; it is flagged once, here, so a reader chasing "issue #157" does not land on a dust-size probe.
This document names real, existing Stellar-ecosystem projects and ties each to
Wayfare capabilities verified in code on 2026-09-25 — not to roadmap
prose. It is a desk-research map from public sources. What it is not, stated
plainly: it is not the answer to backlog #157's actual question. Only asking
a person at one of these organisations settles what they would really need;
nothing in this map substitutes for that outreach
(docs/spike-api-consumers.md reached the same
conclusion on 2026-09-24). Each entry below is a hypothesis with sources, not
a demand signal.
Checked against the tree at 295085b on 2026-09-25. Each row names the
package or endpoint a consumer would actually bind to.
| Capability | Where it lives | Status on 2026-09-25 |
|---|---|---|
| Ladder measurement of a stablecoin → fiat-token corridor, twelve sizes 0.1 → 5000, priced by Horizon strict-send pathfinding | dex/dex.go, route/route.go; GET /api/corridor (docs/api.md) |
Live. Every published quote is kind: "dex" today (route/route.go:186 is the only value reached; docs/spike-adversarial-review.md:261 agrees) |
| Verdicts GOOD/FAIR/POOR/UNUSABLE per size, and the recommendation rule (nothing recommended when nothing is POOR or better) | README "Verdict thresholds" and "The recommendation rule"; route/route.go |
Live, wire contract: recommended present and null when nothing is worth taking |
Integrity states DIRECT / DERIVATIVE (with depends_on) / NO-MARKET / UNKNOWN |
README "Integrity states"; route/route.go |
Live |
| Cross-checked reference mid, two providers, never averaged; >10% divergence publishes no verdict | refrate/cross.go; README "Reference agreement" |
Live |
SEP-1 discovery — an anchor's published stellar.toml, Priceable, and the advertised-SEP inventory via Profile.SEPs() |
anchor/anchor.go |
Live (capability exists in the package; see the API column below for what reaches the wire) |
Counterparty checks: 7 in the default set — AnchorAssetISO4217, SEP10EndpointResponds, SEP24InfoListsAsset, SEP38QuoteServerPublished, IssuerAuthFlags, IssuerFlagImmutability, HomeDomainRoundTrip |
checks/runner.go Default() (checked 2026-09-25); tri-state contract in docs/checks.md |
Live. They appear as the findings block on /api/corridor; they qualify the headline and never move it |
| Historical trend for a corridor, oldest-first, from the hash-chained store | GET /api/corridor/trend (docs/api.md) |
Live, reads stored runs only |
| Browser consumers without a proxy | README "HTTP API": Access-Control-Allow-Origin: *, keyless, read-only |
Live |
| Money as decimal strings on the wire | ADR 006 | Live |
| A live SEP-38 round-trip performed by the project, and an anchor quote published on the wire | sep38/ client; route.KindAnchorSEP38 |
Future. The client and the fee-denomination identity are implemented and unit-tested, including a recorded fixture from testanchor.stellar.org (sep38/live_roundtrip_test.go, checked 2026-09-25), but nothing in production calls the package: route/cost.go:83-87 says an anchor fee "can be wired here" when a SEP-38 corridor is wired in, and a code search of route/ and server/ on 2026-09-25 finds no sep38. call outside tests. No published quote has ever carried kind: "anchor-sep38". Describing anchor quotes as a Wayfare output today would be wrong; this map marks that need future wherever it appears |
| Market-quality metrics (spread, depth, price impact, concentration) on the wire | checks/metric_*.go |
Future — implemented in the tree but unreachable: checks.Runner has no way to run a Metric (README v2 section, checked 2026-09-25; #91) |
Two discrepancies found while verifying the above are recorded in the closing section rather than silently relied on.
What it does today. Freighter is the SDF's non-custodial Stellar wallet, available as a browser extension and a mobile app; users hold their own keys and transact on Stellar, including swaps inside the wallet (freighter.app and docs.freighter.app, both checked 2026-09-25).
What it would need from Wayfare. A per-corridor reading it can render
before or alongside a swap/send: which of the corridors its user is about to
cross is fit to offer at the size being sent, scored against an independent
mid rather than the route's own optimism. Concretely, verified capabilities:
GET /api/corridor for the loss ladder and verdicts at the user's own size
(sizes= accepts arbitrary rungs, docs/api.md); the
recommended: null contract as an honest "don't route through this" signal;
integrity: DERIVATIVE so a wallet can tell a user their naira leg is really
another token's market; and the CORS/decimal-string wire properties that make
a browser extension a natural client (README "HTTP API", checked 2026-09-25).
A swap-interface integration would additionally want anchor-vs-DEX comparison
and depth metrics — future (see the capability table).
Met today? The single-size, verdict-plus-integrity reading is fully met by what ships on 2026-09-25. The swap-comparison reading is future. Whether Freighter would ever surface a third-party corridor-integrity figure in its send flow is unknown — no public statement exists either way; this entry is a hypothesis with a matching API shape, not a demand signal.
Inconclusive — reported as such, deliberately. LOBSTR is a widely used
Stellar wallet (lobstr.co, checked 2026-09-25 for
existence and general purpose only). I could not determine from public
sources whether its send flow exposes any pre-send corridor quality
information, or what data it would need to do so, without speculating about a
product I have not used at integration depth. Rather than invent a "need"
from a marketing page, this entry records: real project, plausible wallet
candidate for backlog #157-style outreach, no verifiable mapping
established.
What it does today. Kotani Pay sells last-mile payout and off-ramp services: businesses pay out in stablecoins (USDC, USDT, cUSD, USDGLO are named on the page) and withdraw to local payment channels across Africa, listing use cases from gig-work payouts to remittances (kotanipay.com/apis, checked 2026-09-25). The page names fiat currencies KES, GHC and XAF among what it connects to.
What it would need from Wayfare. A way to decide, per corridor and size,
whether the on-chain leg its customers will traverse is worth routing through
at all: GET /api/corridor's ladder summary (floor_loss_pct /
worst_loss_pct) plus integrity to catch derivative corridors whose
liquidity is borrowed from another token. If Kotani settles into NGN, GHS or
KES via Stellar fiat tokens, GET /api/corridor/trend adds the "is this
getting worse?" dimension (docs/api.md, checked 2026-09-25). All of
that is live. A per-leg all-in cost figure would additionally want the cost
decomposition and depth metrics — future (see the capability table).
Met today? The route-or-don't-route reading is met today. The
structural caveat, found while checking and reported plainly: Kotani's
public material describes the business without naming Stellar as its
settlement rail, and
https://kotanipay.com/.well-known/stellar.toml returns HTTP 404 (checked
2026-09-25). It therefore appears in this map as a PSP whose corridors
overlap Wayfare's (NGN/GHS/KES off-ramps), not as a verified Stellar
integration; whether it actually moves value over Stellar could not be
determined from its own site. Flagged as the first question outreach should
ask.
What it does today. Eversend operates a USDC off-ramp API: USDC or USDT in, local-currency payouts out via mobile money (M-Pesa, MTN MoMo) and bank rails, with the conversion rate "returned on the quote" and custody via Fireblocks. Its "Networks in" list includes Stellar explicitly, alongside Tron, Ethereum, Solana, Base, Polygon and BNB Chain (eversend.co/platform/apis/usdc-off-ramp, checked 2026-09-25).
What it would need from Wayfare. A benchmark for the rate its own quote
locks: reference_mid with reference_source and reference_pair is a
published, independently checkable mid for exactly the pairs Eversend quotes
(README "Reference agreement", checked 2026-09-25), and the two-provider
divergence fields bound how much that benchmark itself can be trusted.
GET /api/corridor/trend supplies rate history for the same corridor —
relevant when their quote-hold window is short. All live capabilities. Their
own per-quote rate is their business; Wayfare would inform, not replace it.
Met today? Yes for the benchmark-and-trend reading. Caveat: Eversend
has no Stellar issuer presence — https://eversend.co/.well-known/stellar.toml
returns HTTP 404 (checked 2026-09-25) — so Wayfare can measure the corridor
their quotes traverse (e.g. USDC→NGN via NGNC or NGNT), not Eversend's
own product. What Eversend would think of an independent, public benchmark
against its quoted rates is unknown; both the informative and the
awkward readings of that are honest possibilities, and only outreach settles
it.
What it does today. HoneyCoin provides stablecoin payments infrastructure — on-ramps, off-ramps, collections, payouts and FX across 21 African markets through a single API, moving USDC and USDT via mobile money, bank transfer and virtual accounts (honeycoin.app/stablecoin-payments-africa, checked 2026-09-25). The page's own framing is the corridor-loss problem Wayfare measures: "lose another 3-7% in FX spread before the money lands".
What it would need from Wayfare. Independent evidence for exactly that
spread claim, per corridor: the loss ladder and verdicts from
GET /api/corridor measured on mainnet against a published reference mid.
All live capabilities on 2026-09-25.
Met today? The measurement need is met. Caveat, same shape as
Kotani's: no Stellar issuer endpoints were found for HoneyCoin
(https://honeycoin.app/.well-known/stellar.toml returns HTTP 404, checked
2026-09-25) and its public material does not name its chain integrations on
the page checked. Included as a corridor-overlapping PSP, not a verified
Stellar integration.
Negative finding on the PSP category, reported as such: no PSP was found
that publishes its own SEP-1 stellar.toml or operates as a Stellar anchor
in the sense Wayfare's checks/anchor packages would observe. Every PSP
entry above consumes corridor measurements; none could be an anchor/
discovery subject itself on the evidence available. A PSP that also runs an
anchor (with a stellar.toml) may exist in the ecosystem — none was found in
this bounded survey, and that absence is a finding, not a gap in the list to
be filled next pass.
These four are the anchors with African-fiat tokens — the corridor set Wayfare's own case study measures — and all four carry in-repo findings re-verified from primary sources on 2026-09-25. They are the concrete subjects any Wayfare-informs-anchors story has to start from.
What it does today. Issues the naira token NGNC (status live) plus
GHSC and KESC (both pending), all from one issuing account, and publishes
WEB_AUTH_ENDPOINT and TRANSFER_SERVER_SEP0024
(https://ngnc.online/.well-known/stellar.toml,
fetched 2026-09-25; same picture as the in-repo verification at
asset/known.go, checked 2026-09-25).
What it would need from Wayfare. An independent, reproducible reading of
its own corridor: /api/corridor?from=USDC&to=NGNC returns the loss ladder,
verdicts, and the cross-checked USD/NGN reference with both providers named —
a published figure about the market an anchor's token routes through that the
anchor did not produce itself. The findings block adds the issuer-flag
picture (issuer.auth-flags, issuer.auth-immutable) and the declared-SEP
inventory read from their own toml. All live capabilities (checked
2026-09-25).
Met today? Yes — and this is the one case where the interest is not hypothetical: the corridor Wayfare measures IS this anchor's market, and the measurement found its structural floor (25% at dust size; README "What the measurements found", checked 2026-09-25). Whether LINK.IO wants that published is a different question this map does not answer.
What it does today. Issues NGNT, the most widely-held naira token on
mainnet, status live, and declares KYC_SERVER, TRANSFER_SERVER (SEP-6),
WEB_AUTH_ENDPOINT and DIRECT_PAYMENT_SERVER (SEP-31) — but no
ANCHOR_QUOTE_SERVER (https://cowrie.exchange/.well-known/stellar.toml,
fetched 2026-09-25; matches asset/known.go's verification of 2026-08-26,
checked 2026-09-25).
What it would need from Wayfare. As an issuer, the same independent
corridor reading as LINK.IO (live capabilities). As a programmatic anchor
there is a sharper one: Wayfare's anchor.Profile.Priceable and the
SEP38QuoteServerPublished check exist precisely to report that their toml
publishes no machine-readable rate — the fact that keeps NGNT's anchor leg
unpriced. That report is live today; an anchor RFQ comparison would be
future.
Met today? Yes for the discovery/report reading.
What it does today. Issues ZARZ (status live, pegged 1:1 to ZAR, from
an FSCA/SARB-regulated issuer) and USDZ, and is the one African-fiat anchor
in this set that declares ANCHOR_QUOTE_SERVER
(https://zeam.money/.well-known/stellar.toml,
fetched 2026-09-25). However, its quote server's /info still advertises
only USDC, a BRL token and iso4217:BRL — neither ZARZ nor
iso4217:ZAR appears
(https://anchor.zeam.money/sep38/info,
fetched 2026-09-25). This reproduces the finding of
docs/sep38-african-fiat-survey.md (2026-08-30)
with a fresh check.
What it would need from Wayfare. Exactly what the survey doc predicted:
a consumer that treats "declares SEP-38" and "quotes this asset" as different
facts. Today, anchor.Profile.Priceable reports the declared server (live),
and SEP38QuoteServerPublished reports the same as a tri-state check (live).
Checking whether the specific asset is advertised requires fetching the
anchor's /info — sep38/ has the client machinery to do that (implemented
and tested), but nothing in production calls it, so that check is not
something Wayfare runs for any corridor on 2026-09-25 (future as a
published finding; the survey's #183 is the tracked gap). If ZARZ ever
appears in their /info, the full SEP-38 quote comparison (anchor RFQ vs DEX)
is likewise future — see the capability table.
Met today? Partially, and this entry is the map's clearest illustration
of the declared-vs-working boundary: Wayfare's live checks would not
misreport Zeam as priceable-for-ZARZ (they only assert the declared server),
but they would not catch the gap either — the gap is visible in this document
because a human fetched /info on 2026-09-25.
Inconclusive — unchanged from the August survey. Named as an
African-fiat anchor by the Stellar Community Fund's project page and its own
2021 integration post (sources cited in
docs/sep38-african-fiat-survey.md, checked
2026-08-30). Re-checked on 2026-09-25:
https://connect.clickpesa.com/.well-known/stellar.toml does not resolve
(HTTP 000 — DNS failure) and https://clickpesa.com/.well-known/stellar.toml
returns HTTP 404. Whether ClickPesa still operates as a Stellar anchor at all
could not be determined. It stays in the map as an outreach candidate and
a standing unknown, not as a negative: an unreachable domain is a different
fact from a declared-absent quote server, and this document does not collapse
the two.
testanchor.stellar.org (SDF reference implementation) — included for completeness, not as an ecosystem candidate
The SDF's test anchor serves a working SEP-38 API
(https://testanchor.stellar.org/sep38/info
answered with an asset list including stellar:SRT and iso4217:USD, fetched
2026-09-25) and supplied the recorded fixture behind the one verified live
round-trip in sep38/live_roundtrip_test.go (checked 2026-09-25). It is a
test service, not a project Wayfare could inform — listed here because it is
the proof, cited in-repo, that the future anchor-quote capability has at
least one endpoint it demonstrably works against.
Negative finding on the anchor category, reported as such: among the
African-fiat anchors actually reachable on 2026-09-25, zero publish a SEP-38
quote server that advertises the African-fiat asset it issues (ngnc.online:
no server; cowrie: no server; zeam: server present, ZARZ absent). The
consequence for this map is structural: the anchor-uses-Wayfare's-RFQ-story
angle has no live subject in the corridor set Wayfare actually measures
until one of these tomls changes or an unsurveyed anchor appears. The angle
is real (the sep38/ client and route.KindAnchorSEP38 are built and
tested), but on today's evidence it is future for every named anchor.
- Mobile-first African consumer wallets (the e.g. Fonbnk/Accrue class). The 2026-07 Spark+Kotani analysis of mobile-money/stablecoin bridges names API layers in this class (third-party analysis; not checked against their own docs in this spike). They would be natural wallet entries, but no entry was written: without checking a specific product's own documentation for its Stellar exposure and send flow, any "need" assigned to it would be the speculation this issue's constraints forbid. Listed as outreach targets, not mapped entries.
- Non-African corridor anchors (e.g. Montelibero for EURMTL, verified in
asset/known.go, checked 2026-09-25) are measured by Wayfare today and could be informed exactly like LINK.IO or Cowrie. They were left out of the per-project sections to keep this map bounded to the corridor set the project's own case study and prior surveys cover; the pattern transfers unchanged. - Exchanges/DeFi frontends (e.g. Aquarius, Blend) appear in the codebase
only as bridge-asset issuers (
asset/known.go, checked 2026-09-25). No evidence was sought or found that they consume third-party corridor measurements; entering them would have been speculation.
Reported plainly, per the issue's own acceptance criteria:
- The actual demand question is untouched. This map names projects and
matches verified capabilities; it does not know whether any of them want
what Wayfare publishes. Backlog
#157(GitHub #217) was closed inconclusive on 2026-09-24 for exactly this reason, and this document supplies its input, not its answer. Every "what it would need" above is a hypothesis for outreach, with the project list here as the call sheet. - Three PSPs are corridor-overlapping but not verified Stellar
integrations. Kotani Pay, Eversend and HoneyCoin all do stablecoin
off-ramping into the fiat currencies Wayfare measures, and Eversend lists
Stellar among its input networks — but none publishes a
stellar.toml, and Kotani's and HoneyCoin's use of Stellar specifically could not be confirmed from their own sites (checks dated 2026-09-25). - The anchor-quote angle has no live subject. No reachable African-fiat
anchor publishes a SEP-38 server advertising its own fiat asset (see the
negative finding above; checks 2026-09-25). Any "anchor compares Wayfare's
DEX reading to its own RFQ quote" story is future twice over: the
Wayfare side is not wired (
route/cost.go:83-87), and no named anchor currently offers the other side of the trade. - ClickPesa remains undetermined — DNS/404 on both candidate toml domains, 2026-09-25, consistent with the 2026-08-30 survey.
- LOBSTR is an unmapped wallet — real, plausible, but no verifiable mapping was established without deeper product inspection; recorded rather than padded.
- Two repository discrepancies were found while verifying and are flagged,
not fixed here (this is a docs-only issue):
- The README "Verification status" table states the live SEP-38
round-trip is "Verified — recorded fixture from testanchor.stellar.org
in
sep38/testdata/live/", andsep38/live_roundtrip_test.gosupports that (both checked 2026-09-25) — but the README's own "HTTP API" section still says "a live SEP-38 round-trip has never been performed — see #180" (checked 2026-09-25), and #180 is closed. The two README passages contradict each other; the table's newer claim is the one the code supports. - The README's v2 section says "three [counterparty checks] run per
corridor";
checks/runner.goDefault()(checked 2026-09-25) defines seven. The code is the stronger evidence; the README count is stale.
- The README "Verification status" table states the live SEP-38
round-trip is "Verified — recorded fixture from testanchor.stellar.org
in
Wayfare repository, all checked 2026-09-25 at commit 295085b:
README.md (Non-goals, HTTP API, Verification status, Roadmap),
CONTRIBUTING.md (Invariants), docs/api.md, docs/non-goals.md,
docs/checks.md, docs/about.md, docs/backlog.md (#157, #273),
docs/spike-api-consumers.md, docs/spike-sdk-surface.md,
docs/sep38-african-fiat-survey.md, anchor/anchor.go, dex/dex.go,
refrate/refrate.go, refrate/cross.go, sep38/sep38.go,
sep38/live_roundtrip_test.go, route/route.go, route/cost.go,
checks/runner.go, checks/sep38_quote_server.go, asset/known.go.
GitHub API (Wayfare-labs/wayfare), checked 2026-09-25: issues #326, #157, #180, #217 and their comments.
External, each with its check date:
| Source | Checked |
|---|---|
| freighter.app / docs.freighter.app | 2026-09-25 |
| kotanipay.com/apis | 2026-09-25 |
| eversend.co/platform/apis/usdc-off-ramp | 2026-09-25 |
| honeycoin.app/stablecoin-payments-africa | 2026-09-25 |
| lobstr.co (existence and general purpose only) | 2026-09-25 |
| ngnc.online/.well-known/stellar.toml | 2026-09-25 |
| cowrie.exchange/.well-known/stellar.toml | 2026-09-25 |
| zeam.money/.well-known/stellar.toml | 2026-09-25 |
| anchor.zeam.money/sep38/info | 2026-09-25 |
| testanchor.stellar.org/sep38/info | 2026-09-25 |
connect.clickpesa.com and clickpesa.com stellar.toml (DNS failure / HTTP 404) |
2026-09-25 |
kotanipay.com, eversend.co, honeycoin.app /.well-known/stellar.toml (HTTP 404) |
2026-09-25 |
| spark.money — Africa's Mobile Money Networks Meet Stablecoins (third-party analysis; outreach-target names only) | 2026-09-25 |
- Issue #326 — this task
- Backlog
#157/ GitHub #217 — the consumer spike this map is input to (docs/spike-api-consumers.md) - docs/sep38-african-fiat-survey.md — the anchor survey whose findings this document re-verified on 2026-09-25
- docs/spike-sdk-surface.md — what a client library would need to expose, once consumers exist
- docs/about.md — the audience statement this map tests against named projects