Two Doors, One Corpus: How Project Hermes Redesigned Its Front Page Around the People Who Actually Use It
Hermes v0.7.1 and v0.7.3 replaced the old intake-form-as-homepage with a two-door landing — one door for witnesses, one for researchers — and a mission statement that reframes the entire relationship: Hermes does not grade the witness. Hermes helps the witness make better data.
For decades, the front door of a UFO reporting site has been a form. You arrive, and immediately you are confronted with fields: latitude, longitude, elevation, date-time in UTC. The message is clear, even if nobody writes it on the page: prove you belong here. If you cannot enter your sighting in the vocabulary of the site, the site does not want your sighting.
Project Hermes just stopped doing that.
The two-door landing
The new root of projecthermes.tech (v0.7.1) is no longer the intake form. The form has been moved to /intake, where it still lives, still works, and is still linked from the front page. The new root is a threshold with two parallel entry points.
On the left: "Report what you saw." This is the witness door. It leads to the conversational intake (more on that in the next post), and its copy is direct: you don't need to know what it was, you don't need to be certain, you only need to describe what happened, in your own words, as honestly as you can remember.
On the right: "Examine the system." This is the researcher door. It exposes the corpus size, the current methodology version, the integrity hash, and the date range, live-read from /api/corpus at page load. A researcher can see, before clicking anything, whether the system is in a state they want to cite.
Both doors share the same corpus, methodology, rule catalog, and audit trail. They only differ in how the visitor is oriented in the first three seconds. That is a small design choice with a large consequence: it means a frazzled witness at 2 a.m. and a PhD candidate building a literature review no longer start on the same confusing page.
The scope statement, moved to the front
Below the doors sits a "What Hermes is / what Hermes is not" contract. It states in plain language that confidence scores measure thoroughness of checking, not likelihood of any interpretation. It states that witness-reported kinematics — hovering, instantaneous turns, impossible accelerations — are preserved verbatim in the case record, but are not treated as instrumented measurements. It states that Hermes is not a prediction system and does not describe "UAP activity," only reporting conditions and report volume.
This language was already present in Hermes — in API disclaimers, in methodology docs. What changed in v0.7.1 is that it now appears on the landing page. You cannot miss it. You cannot fail to encounter it before you submit a report or cite a case. Surfacing epistemic scope before the first click is, in a small way, a research ethic.
The mission statement (v0.7.3)
v0.7.3 added one more piece, placed between the header and the two doors so both audiences read it before choosing a path:
"Hermes is here to help. We really want your data, and we bet you're here because you had an experience that left you with some questions. No, we don't have answers — but we really want your data, and we want to value your data. We're here to help you make better data, so that together we can formulate better questions."
Read that twice. It is not promising to explain the phenomenon. It is promising to take the witness's contribution seriously, improve it procedurally, and turn it into something a scientist can work with. Feedback on submissions — present and future — will be affirming of the contribution and forward-looking about technique, never scolding about what was already done. That is a posture decision, and in UAP research, posture is rare.
Why it matters for the study of UAP
The single largest bottleneck on UAP data quality is not instruments. It is the fact that witnesses, historically, have been treated as suspects by some communities and as true-believers by others, and almost never as collaborators. The two-door design is infrastructure for collaboration: the witness door is not a simplification of the researcher door, and the researcher door is not a gate against the witness door. They coexist, and the corpus they both feed is the same corpus.
If more UAP platforms adopted this posture — that witnesses and researchers have different needs and both deserve first-class treatment — the field would be in a materially different place five years from now. This is one platform. But the example is on the record.
Visit the new front page at projecthermes.tech.
Project Hermes and UFO Index are affiliated projects.