The UFO archive that investigates itself.
We pulled in every public UFO/UAP source we could find, turned it into a graph, and put autonomous Watchers on the case. They build the wiki, read new government releases as they drop, and keep investigating after the headline cycle moves on.
Agents on Telegram · Discord · X · token-holder web UI
First we collect.
Then we connect.
Open Sky starts with the public UFO/UAP corpus: government releases, FOIA files, case databases, witness reports, photos, videos, transcripts, and old archives. The system breaks them into people, places, dates, craft descriptions, sensor events, claims, sources, and citations.
That graph gives the Watchers context. A new file can be compared against older cases, launch windows, weather, astronomy, aircraft, satellites, and every similar claim we have already indexed.
Day Zero — new file appears → archived, linked, investigated
Government UFO files do not arrive as a clean database. They show up as PDFs, scans, videos, photos, tables, redactions, bad OCR, and source pages that disappear later. Open Sky is built to catch the release, preserve the file, read it, and attach it to the graph before the trail goes cold.
We are already running on Project Blue Book, AARO material, FAA records, NASA’s UAP study, historical archives, witness databases, and 197 tracked source lanes. When new files come out, the Watchers OCR them, extract names, dates, locations, objects, sensors, and claims, then push the work into the wiki for continuous investigation.
Files become investigations. The AI has to show its work.
Every claim starts at the file.
Each page points back to the document, photo, video, witness report, or government release that made the claim. If we cannot trace it, we say so.
The archive remembers connections.
Names, dates, locations, units, aircraft, sensors, shapes, and explanations become linked records instead of loose notes buried in PDFs.
The Watchers keep working.
Agents keep checking weather, satellites, launches, aircraft, astronomy, and older cases. Unknown means unresolved, not magic.
Agents that read the file. Then check the sky.
Watchers ingest releases, extract claims, run weather, launch, astronomy, aircraft, and satellite checks, then flag what still needs a human investigator.
- Ingest
- Extract
- Check
- Audit
197 trusted sources. One connected corpus.
We are not waiting for a perfect archive to appear. We are pulling the scattered record together now, then letting the graph and wiki get stronger every time a new release, witness file, or source lane lands.
- 001Project Blue Book
- 002AARO Historical Record
- 003FAA FOIA
- 004NASA UAP Study
- 005NUFORC
- 006MUFON
- 007NICAP Archive
- 008Condon Report
- +189 more sources
$OSP funds the archive. The public trail stays open.
$OSP coordinates source lanes, Watcher runs, deeper case work, and the investigator UI without turning the archive into hype.
The landing version stays simple: public evidence, holder access, and accountable funding for the agents that keep reading.
- Symbol
- $OSP
- Network
- Coming soon
- Purpose
- Archive funding + governance
- Access
- Investigator UI + agent layer
We are not asking you to believe.
We are asking for the file.
Then we open the trail.
The archive preserves the strange parts and the boring parts. Every claim should point back to a source, every explanation should show its work, and every unresolved case should stay open.
If the files are public, the public should be able to investigate them.