Draft — not legal advice, and not yet reviewed by counsel. This is an engineering draft written to be factually accurate about how the system behaves. It must be reviewed by a qualified lawyer before it is presented to users as binding terms. See Review status.
Last updated: 2026-08-04
Stellar Intel is an information and routing layer: it compares off-ramp rates across Stellar anchors, publishes an anchor reputation oracle, and exposes that data through a web interface, a public API, and an agent surface.
It is not a money transmitter, an exchange, a broker, or a custodian. It does not move your money. Every transfer is between you and a third-party anchor.
Stellar Intel is non-custodial by construction, not by promise — there is no
code path through which it could take custody
(docs/NON_CUSTODY.md).
Specifically, Stellar Intel never holds:
You sign every transaction. Nothing moves without a signature you produce in your own wallet.
This is the most important thing on this page.
Most rates shown are indicative: derived from anchors' published fee schedules and live exchange rates. They are an estimate of what you would receive, not a binding quote, and the amount you actually receive may differ.
Where an anchor supports SEP-38 firm quotes, a quote you obtain during execution is binding on that anchor for the period it states — and it expires. Rates carrying a firm quote are labelled distinctly in the interface from those that are indicative.
Stellar Intel does not set, guarantee, or underwrite any rate. The rate is the anchor's.
Anchor scores are computed from observed outcomes and availability probes, using
the published methodology in
docs/ANCHOR_REPUTATION.md. Scores derived from a small
number of observations are labelled as such.
A score is a description of past behaviour. It is not a recommendation, a guarantee of future performance, a statement about an anchor's solvency or licensing, or a substitute for your own diligence.
Listing an anchor is not an endorsement of it.
Anchors are independent third parties. Stellar Intel does not control them and is not responsible for their conduct, solvency, availability, or compliance.
Anchor data is fetched from the anchors themselves and may be stale, incomplete, or wrong. An anchor may change its fees, suspend a corridor, or fail to settle, without notice to us.
The service is provided as is, with no guarantee of availability, accuracy, or fitness for a particular purpose. It depends on third-party infrastructure — anchors, the Stellar network, and data providers — any of which may be unavailable.
Data may be cached. The interface indicates freshness where it is known.
To the maximum extent permitted by law, Stellar Intel and its contributors are not liable for losses arising from use of the service, including losses from an inaccurate rate, an anchor's conduct, a failed or delayed settlement, or a transaction you signed.
Counsel must review this clause. Enforceability of a liability limitation varies significantly by jurisdiction, and some consumer-protection regimes void it outright.
Stellar Intel's regulatory posture rests on never taking custody. The architectural
basis is recorded in docs/JURISDICTIONAL.md, which is a memo
for reviewers, not a country-by-country legal opinion.
The service is not offered to users in jurisdictions where it would be unlawful, and it is not directed at any jurisdiction in particular. If use is restricted where you are, do not use it.
Vulnerability reporting is covered by docs/SECURITY.md. The
in-scope and out-of-scope surfaces are in
docs/THREAT_MODEL.md. Notably, your wallet and the anchors'
systems are outside Stellar Intel's control and therefore outside its threat model.
These terms may change. Material changes will be reflected in the "Last updated" date. Continued use after a change constitutes acceptance.
The user-facing summary rendered in the product (issue #739). It must remain consistent with the sections above; if one changes, change both.
Stellar Intel is non-custodial. You sign every transaction with your own wallet. Rates are live quotes, not guarantees.
Kept to three sentences deliberately — it appears in a banner, and a disclaimer too long to read is not a disclaimer. Each sentence maps to a section above: non-custody (§2), user-signed (§2), rates are estimates (§3).
components/offramp/DisclaimerBanner.tsx renders this text and links here, so
this document is the single source for the wording.
| Item | Status |
|---|---|
Factually consistent with NON_CUSTODY.md |
Yes |
Factually consistent with JURISDICTIONAL.md |
Yes |
Factually consistent with THREAT_MODEL.md |
Yes |
| Reflects that most rates are indicative (#720) | Yes |
| Reviewed by a qualified lawyer | No — required before publish |
Two things a reviewer should look at first: the liability limitation in §8, and whether §9's approach to jurisdiction is sufficient for the corridors actually served.
docs/NON_CUSTODY.md — why custody is structurally impossibledocs/JURISDICTIONAL.md — regulatory posturedocs/THREAT_MODEL.md — what is in and out of scopedocs/ANCHOR_REPUTATION.md — how scores are computeddocs/SEP_COMPLIANCE.md — which anchors support firm quotes