Overview
On 10 July 2026, the U.S. Department of War published the fourth tranche of records under the Presidential Unsealing and Reporting System for UAP Encounters (PURSUE). The Department described the release as part of the rolling government-wide effort to locate, review, declassify and publish unresolved UAP-related material following the presidential transparency directive issued earlier in the year. Department of War — Release 04 announcement Department of War — PURSUE Release 04 portal
Release 04 contained 40 new files: 14 documents, 19 videos, four audio recordings and three images, drawn from the Department of War, NASA, CIA, FBI and Department of Energy. The official portal made 227 MB of documents and 1.4 GB of video available for bulk download. Department of War — PURSUE archive CBS News — 10 July 2026 release coverage
The tranche is unusually useful because it places several very different kinds of evidence beside one another.
There are modern military infrared videos submitted to AARO, including 2025 recordings from the Yellow Sea and East China Sea; Navy Range Fouler records from the Atlantic and eastern United States; a 2015 Department of Energy incident report from the Pantex nuclear-weapons facility; a 1949 classified conference at Los Alamos concerning green fireballs; early Air Force intelligence studies; Project Blue Book administrative material; and three NASA images from the 1996 STS-80 Space Shuttle mission. Department of War — Release 04 record index
The most publicly discussed file was a 32-second 2020 Atlantic infrared video, DOW-UAP-PR116, quickly nicknamed the “floating brain” or “jellyfish” because of its irregular silhouette. The official record is much less dramatic than some viral versions. The paired Navy debrief says the object travelled with the wind, did not manoeuvre or change direction and appeared to the observer like a large, somewhat deformed balloon. AARO nevertheless released it as unresolved rather than assigning a definitive identification. AARO — UAP Report Documents DVIDS — DOW-UAP-PR116
That combination—visually suggestive imagery, comparatively mundane reported behaviour and unresolved official status—captures the central lesson of Release 04.
Unresolved is not a positive classification of extraordinary technology.
PURSUE itself defines an unresolved case as one for which the government cannot make a definitive determination about the observed phenomenon, including cases where the available data are simply insufficient. Department of War — PURSUE initiative
The fourth release is therefore more valuable as an evidence archive than as a single disclosure claim.
Chronology
10 July 2026 — PURSUE Release 04.
The fourth tranche publishes 40 records from five agencies.
Release inventory
The Department's release announcement is brief. It states that the fourth tranche had been published and that additional releases would follow.
The portal provides the real substance.
The 40 files include material from five agencies. Contemporary reporting counted 28 Department of War files, seven NASA files, two CIA records, two Department of Energy records and one FBI record. PURSUE Release 4 file index
The Department of War material is itself heterogeneous.
Some files are modern AARO case media.
Others are historical Air Force or intelligence documents transferred into the PURSUE archive.
Several modern videos are paired with narrative reports, which considerably increases their analytical value compared with isolated clips.
The NASA material includes images associated with STS-80 and audio material from the agency's historical holdings.
The DOE records include both the 1949 Los Alamos green-fireball conference and the 2015 Pantex security incident.
This is why the release should not be treated as though all 40 files represent 40 UAP cases of equal quality.
A historical conference transcript, a modern infrared video, an interview audio file and a Space Shuttle image require different methods of assessment.
The common feature is archival relevance, not evidential uniformity.
Unresolved status
PURSUE's own definition should govern interpretation of the tranche.
The Department says the archive contains unresolved material for which the government is unable to make a definitive determination about the nature of the observed phenomena. It explicitly states that this can occur for several reasons, including insufficient data, and says resolved cases continue to be reported separately under AARO's statutory obligations. Department of War — PURSUE initiative
This matters because the word can easily be given a stronger meaning than the government assigns to it. A case may remain unresolved because a genuinely unusual object was recorded by multiple independent sensors and resists known explanations. Another may remain unresolved because it consists of only a few seconds of low-resolution video without range or contextual metadata. Both can appear in the same archive. The evidential weight is not the same. Release 04 contains examples across that spectrum, and AARO's own descriptions are often careful to avoid turning visual impressions into analytical conclusions.
Atlantic video
DOW-UAP-PR116 is the clearest example of how public interpretation can diverge from the underlying report. The government says U.S. Northern Command submitted 32 seconds of infrared footage from a U.S. military platform concerning a 2020 Atlantic Ocean event. The paired Range Fouler Debrief, DOW-UAP-D091, contains the observer's description.
The object was described as darker or maroonish, approximately 12 to 15 feet in height, travelling with the wind and showing no manoeuvre or change in direction. The observer wrote that its structure appeared similar to a large, somewhat deformed balloon, while also stating that the crew was unable to verify that interpretation during the pass. AARO — UAP Report Documents DVIDS — DOW-UAP-PR116
This is important primary evidence. The apparent shape seen in the infrared footage is irregular enough to invite comparison with a biological organism or multi-part structure. The original witness account does not describe biological behaviour, controlled turns, rapid acceleration or reaction to the aircraft. It describes an object whose motion appeared broadly consistent with atmospheric drift. A balloon or balloon cluster is therefore a plausible conventional hypothesis. It is not a formally established government resolution. AARO's public designation remains unresolved. That distinction is exactly where the evidence currently sits.
Paired debrief
Without DOW-UAP-D091, the 32-second video would be substantially more ambiguous. A low-resolution infrared silhouette does not reveal scale. It does not reveal true colour. It does not automatically reveal material. It does not distinguish a single rigid object from a cluster of lightweight objects whose thermal contrast merges at the sensor's resolution. The debrief adds independent observational context. The aircrew estimated height. They perceived maroon coloration. They observed wind-correlated movement. They reported no manoeuvring. They compared the structure to a deformed balloon. None of those observations is perfectly measured.
AARO explicitly warns that descriptive and estimative language in such reports reflects the reporter's subjective interpretation and should not be treated as conclusive evidence of intrinsic object characteristics. That qualification cuts in both directions. The balloon description is not proof that it was a balloon. The unusual silhouette is not proof that it was a structured craft. The paired record narrows interpretation while leaving uncertainty.
AI enhancement
Within days of the release, enhanced versions and stills of the Atlantic footage circulated widely online. Contemporary reporting explicitly noted that an AI-enhanced version had gone viral. EarthSky — fourth PURSUE release and “floating brain” video
This creates an evidential problem independent of the original sighting. Modern generative enhancement can enlarge low-resolution imagery and produce visually coherent edges, textures and intermediate frames. It does not recover information that was never recorded by the original sensor. Where a model predicts detail, the resulting image can look more informative while actually becoming less suitable as evidence of physical morphology. The distinction is especially important for UAP footage. A generated edge may be mistaken for an appendage. Interpolated motion may appear smoother or more articulated than the source. A sharpened thermal blob can acquire a surface structure that exists only in the enhancement model. For Release 04, the evidential hierarchy should therefore remain simple:
the original government video and paired debrief are the primary record; ordinary non-generative processing can be useful if its transformation is documented; AI-generated enhancement should be treated as an interpretive visualisation rather than recovered sensor evidence. The viral attention surrounding the Atlantic clip makes this one of the most useful modern examples of why provenance must extend not only to the original file but to every processing step applied after release.
Range Fouler case
Another notable Release 04 case is DOW-UAP-PR112.
The U.S. Navy submitted 20 seconds of infrared footage recorded from a civilian aircraft in 2019. The paired Range Fouler Debrief, DOW-UAP-D090, contains a striking statement from an aviator who reported 28 years of service with the Air Force and Navy.
The observer said the object displayed flight characteristics unlike anything he had seen during that period of service and described a small object travelling in the opposite direction at apparently high speed. The observer said it was tracked visually for approximately 10 to 15 seconds before recording began and that increasing zoom caused the object to leave the sensor field of view. DVIDS — DOW-UAP-PR112
The testimony deserves weight. An experienced military aviator has relevant knowledge of aircraft behaviour and operational airspace. It does not supply missing range. The public video shows an area of contrast and a sensor change in zoom/display mode before the object leaves the frame. AARO's description does not state that physical velocity was calculated. That difference matters. An observer can accurately report that an object appeared extraordinarily fast without possessing the information necessary to convert angular movement into a physical speed. The case is therefore more interesting than a wholly context-free clip, but the public file does not independently validate extraordinary kinematics.
Witness experience
The 2019 debrief is a good illustration of a recurring problem in UAP analysis. Professional experience can improve recognition. A senior aviator is less likely than an untrained observer to confuse a familiar aircraft type under routine conditions. But range ambiguity affects experienced observers too. A small nearby object moving moderately can sweep rapidly across a field of view. A larger distant object moving quickly can produce similar angular motion. If the observer cannot establish distance, apparent speed remains uncertain. The correct use of the witness's 28 years of experience is therefore as evidence that the event was unusual to that observer in that operational context. It should not be converted automatically into a quantitative measurement that the source does not contain.
Yellow Sea
DOW-UAP-PR104 consists of 18 seconds of infrared footage submitted by U.S. Indo-Pacific Command. AARO's neutral video description says the sensor pans to track an area of contrast resembling a six-pointed star, keeping it generally centred on screen. DVIDS — DOW-UAP-PR104 The wording is deliberately visual. AARO does not say a six-pointed physical craft was observed. It says the displayed area of contrast resembles that shape. That distinction is important because point sources and small unresolved objects can acquire multi-lobed shapes through optics, diffraction, focus, saturation, image processing or sensor aperture geometry. A six-point pattern in the image plane is therefore not sufficient to infer a six-point physical body. The public clip also does not provide enough information to establish distance, scale or physical velocity.
Its value lies in preserving a military sensor observation that remains unattributed. Its morphology should remain an image description rather than a vehicle classification.
East China Sea
DOW-UAP-PR105 is substantially longer.
It contains approximately five minutes of infrared footage submitted by Indo-Pacific Command from a 2025 East China Sea event. AARO's timeline says the sensor pans to track an area of contrast, changes settings, zooms and later loses useful content. Portions of the target intermittently lose distinctiveness against the background. AARO — UAP Report Documents
Long duration increases potential analytical value. It provides more frames from which apparent motion, sensor response and tracking behaviour can be examined. It still does not create information that is absent from the release. The public record does not provide a complete range solution, platform trajectory or enough calibrated metadata to reconstruct the object's three-dimensional motion independently. Nor does AARO's descriptive text claim extraordinary manoeuvring. The object remains unresolved because a definitive attribution has not been made. That is different from a government finding that five minutes of anomalous flight were demonstrated.
Sensor behaviour
Several Release 04 videos make an important methodological point visible: the sensor itself changes during the recording. Zoom changes. Contrast changes. Display modes change. Tracking reticles appear or disappear. Targets can lose distinctiveness against the background.
DOW-UAP-PR115, an eight-second 2019 Air Force video, is especially useful because AARO adds a technical comment. The target visually flickers, and AARO explains that infrared targets with temperatures similar to the surrounding environment can blend into the background or appear to flicker because of the sensor system's auto-gain control filters. AARO — UAP Report Documents
This does not identify the object. It constrains interpretation of the imagery. A disappearing thermal signature is not necessarily a physical object vanishing. Apparent brightening is not necessarily an energy emission. Changing morphology is not necessarily shape change. The sensor pipeline must be modelled before a physical interpretation is assigned. Release 04 is valuable partly because AARO increasingly publishes these caveats alongside the imagery.
Historical custody
DOW-UAP-PR113 is a 1996 video transferred by the Navy UAP Task Force to AARO in 2022.
AARO states that no formal UAP data-handling practices existed when the media was originally reported to the Task Force and that the file had been digitally altered before being reported. It is presented publicly as AARO received it. AARO — UAP Report Documents
The released video contains a short transit followed by repeated frame-by-frame playback, slow-motion replay and a held still frame. This is not evidence of deception. Those alterations may simply represent an earlier analyst or observer trying to examine a brief event. The problem is scientific provenance. If the native recording is absent, it becomes difficult to determine which properties belong to the original sensor and which were introduced during later editing. This limits claims about frame timing, speed and image morphology. The case is a strong demonstration of why AARO's modern reporting and workshop papers place so much emphasis on preserving native sensor files. Historical evidence can remain useful while being permanently constrained by earlier handling.
Evidence classes
The modern military footage attracted the greatest attention, but some of the historically strongest material in the tranche is documentary. The Department of Energy released DOE-UAP-D004, a transcript of a 16 February 1949 Conference on Aerial Phenomena at Los Alamos Scientific Laboratory.
Participants included military personnel, federal security officials, meteor expert Lincoln LaPaz and prominent Los Alamos scientists including Norris Bradbury, Edward Teller and Frederick Reines. The conference concerned reports of green fireballs and other aerial phenomena observed in New Mexico. PURSUE — DOE-UAP-D004 record archival transcription of DOE-UAP-D004
This is significant primary evidence of government concern. The meeting was classified Secret. Officials were attempting to determine whether the observations were meteoritic, technological or otherwise explicable. LaPaz argued that aspects of the reports differed from ordinary meteorite phenomena, while other participants explored electrical and physical possibilities. The correct historical conclusion is not that the conference established extraordinary vehicles. It is that senior scientists and security officials considered the reports sufficiently unusual and relevant to a nuclear laboratory to warrant a classified scientific discussion. That distinction makes the document valuable without requiring an extraordinary interpretation.
Green fireballs
The Los Alamos material inevitably raises the recurring question of UAP reports near nuclear facilities. The geographic association is real within the historical record. The inferential meaning is much harder. Nuclear sites are strategically important, heavily guarded and observed by personnel trained to report unusual activity. They therefore create a strong collection environment. That can produce a larger documentary record even if the underlying rate of unusual phenomena is not uniquely high. Conversely, security significance is precisely why repeated unusual observations near such sites should not be ignored. The scientific question is comparative. Were unexplained observations more frequent around nuclear sites after accounting for observation density, reporting practice and military activity?
The 1949 transcript cannot answer that question on its own. It provides one unusually rich historical dataset for asking it.
Pantex
Release 04 also includes a much more recent Department of Energy record: DOE-UAP-D005, concerning a 1 September 2015 incident at the Pantex Plant near Amarillo, Texas. Pantex is a highly sensitive U.S. nuclear-weapons facility.
The official PURSUE description says the file contains imagery and a report documenting an unidentified object intruding the airspace above the plant. It also notes that two pages had previously been released in more redacted form in Release 02. PURSUE — Pantex record
The released report says the plant's ground surveillance radar detected an unknown object moving slowly on a northerly trajectory and that protective-force personnel responded under an existing unmanned-aerial-system incident plan. Ground personnel visually confirmed the object and followed it for several miles. The record describes slow movement, no obvious sound and no immediately identifiable propulsion, with witness descriptions varying. Pantex report archive copy sourced from the official release
This is a stronger evidential situation than a single unsupported anecdote. There was institutional detection, a security response and direct observation. It still does not establish extraordinary technology. The reported speed—around 10 to 15 mph—is not anomalous in itself.
The object's appearance and lack of visible propulsion are interesting because they resisted immediate identification, but a slow airborne object near a facility has conventional possibilities including balloons and unmanned systems.
The significance is that a real security intrusion occurred and remained unidentified in the released record.
Nuclear association
The presence of both Los Alamos and Pantex records in the same tranche creates a powerful narrative temptation. Historical green fireballs appeared near the atomic-weapons complex. A later unidentified object entered Pantex airspace. Therefore, one might infer a persistent external intelligence monitoring nuclear weapons. The documents do not establish that chain. They show separate unresolved observations in environments where unidentified airborne objects matter greatly. A balloon entering ordinary countryside may be harmless. The same object entering protected nuclear airspace can trigger a security response and generate a detailed government record. This creates a selection effect. Sensitive sites preserve incidents that might never be documented elsewhere. That does not make the events unimportant.
It means the location cannot be treated as evidence of intent without additional behavioural information. Release 04 provides useful nuclear-site records. It does not establish a common actor connecting 1949 and 2015.
Air Force records
The tranche also contains DOW-UAP-D093 and DOW-UAP-D094, analyses of flying-object incidents from 1948 and 1949, together with DOW-UAP-D097, a 1948 Project Sign progress report. These documents belong to the formative period of U.S. military UFO investigation. They are valuable because they preserve what intelligence officers were considering before Project Blue Book became the long-running public programme. The documents should be read historically. An intelligence report can record extraordinary claims, proposed hypotheses and unresolved observations without representing the final institutional position of the Air Force. The fact that officials considered foreign technology, unusual aircraft or other possibilities demonstrates uncertainty. It does not establish which explanation was correct.
The Release 04 archive is especially useful here because it allows modern researchers to compare documents that previously existed across separate archives and FOIA collections within one indexed programme.
Blue Book material
DOW-UAP-D092 contains records concerning the Department of the Air Force committee reviewing Project Blue Book in 1966-67. This overlaps with the institutional history covered in the UAPRAD deep dive on the 1966 congressional hearing and the road to the Colorado Study. Its inclusion in Release 04 is important for provenance and accessibility rather than because it changes the historical conclusion by itself. The records document the Air Force's concern with Blue Book staffing, scientific review, university involvement and public handling during the period that led into the Condon project. Their value lies in allowing readers to inspect the bureaucratic process directly.
PURSUE therefore functions not only as a source of new sensor clips but as a republishing mechanism for government records that illuminate how earlier UAP policy was made.
STS-80 images
NASA contributed three images labelled as unidentified-object imagery from the STS-80 Space Shuttle Columbia mission in 1996. The official PURSUE portal describes them as photographs of an unidentified object in low-Earth orbit. Department of War — Release 04 portal The word unidentified needs the same care here as elsewhere.
Space Shuttle imagery contains numerous possible sources of bright or unusual objects: ice particles, debris, payload-related material, distant satellites, optical effects and other objects whose distance may be difficult to determine from one camera perspective.
The presence of the images in an official UAP archive establishes that the government considered them responsive to the release directive. It does not establish an exotic object in orbit. A scientifically persuasive interpretation would require the original sequence, camera geometry, shuttle attitude, lighting conditions and contemporaneous mission context rather than isolated stills. The three images are therefore better regarded as historical NASA evidence records than as resolved physical cases.
Release architecture
PURSUE places material from different eras and agencies into a single visual interface. This improves discoverability. It can also flatten distinctions that matter. A CIA memorandum may reproduce a witness story. A Navy Range Fouler form may record a first-hand military observation. A NASA image may be an original sensor frame. A later digitally edited UAPTF video may be a derivative file with incomplete custody history. All can appear as neighbouring cards. The archive interface therefore should not be treated as an evidential ranking. Government custody authenticates the file's place in the government record. It does not make every file equally direct evidence of the physical event it describes. This is one of the most important lessons of all the PURSUE tranches and is particularly visible in Release 04.
What the release improves
Release 04 materially improves public UAP research in several ways. First, it pairs some imagery with narrative context. The Atlantic PR116/D091 pairing is much more useful than the clip alone. Second, AARO often supplies explicit sensor descriptions. The office notes zoom changes, tracking behaviour, auto-gain effects and prior digital alteration. Third, the historical documents are indexed by agency, incident date and location rather than buried in disconnected archival holdings. Fourth, the release allows public analysts to distinguish what government investigators actually wrote from later retellings. These are genuine transparency gains. The scientific gains remain more uneven. Many videos still lack the full telemetry required for quantitative reconstruction. Redactions remove operational context.
Some historical media are derivative. Some records document witness estimates rather than instrument measurements. Transparency therefore increases the ability to evaluate uncertainty without automatically eliminating it.
What remains unestablished
Release 04 does not demonstrate one common UAP phenomenon. The Atlantic object appears behaviourally compatible with wind drift. The eastern United States case contains stronger witness language but lacks public range data. Yellow Sea and East China Sea footage show unresolved areas of contrast whose true physical morphology is uncertain. Pantex documents an important security intrusion without identifying the object. Los Alamos documents serious historical scientific concern without resolving the green-fireball phenomenon into extraordinary technology. The STS-80 images remain context-limited. The evidence classes differ too much to support a single explanatory conclusion. The tranche is therefore scientifically heterogeneous by design. The appropriate question is not “what does Release 04 prove?”
It is “what does each file allow us to test?”
Historical significance
The fourth PURSUE tranche illustrates a mature phase of the 2026 transparency programme. The early releases established that the government would publish large quantities of material. Release 04 increasingly shows the value of linking records together. Video plus debrief is better than video alone. Historical transcript plus institutional provenance is better than a later anecdote. Sensor description plus imagery is better than a viral clip stripped of context. The tranche also exposes a modern problem that earlier UFO archives did not face at the same scale: generative manipulation. In the analogue era, investigators worried about copied photographs, retouching and missing negatives. In 2026, a government video can be downloaded, generatively enhanced and redistributed to millions of viewers within hours.
The altered version may look clearer than the original precisely because the model has invented a visually coherent answer to uncertainty. That makes provenance part of public scientific literacy. Researchers need to preserve hashes, original files, release metadata and processing history. They need to label visualisations as visualisations. And they need to return to the original record before making claims about shape, structure or motion. Release 04 also reinforces a broader historical pattern. Government UAP archives contain both events and responses to events. The Los Alamos transcript tells us how officials reacted to green fireballs. The Pantex record tells us how a nuclear facility responded to an airspace intrusion. The Range Fouler forms tell us what military crews perceived. The videos tell us what a sensor displayed.
Those layers are related, but none is interchangeable with another.
As of 26 August 2026, Release 04 is best understood as a substantial multi-agency evidence release whose strongest contribution is not proof of one phenomenon but improved access to the different records needed to distinguish one claim from another.
Evidence assessment
PURSUE Release 04 is one of the stronger tranches as an archival package because several records contain enough contextual material to expose both the interest of the observations and the limits of the evidence. The Atlantic “floating brain” case is the clearest example. The imagery looks unusual. The observer's own debrief describes wind-correlated motion, no manoeuvring and a balloon-like structure. A balloon or balloon cluster is therefore a plausible explanation. The case remains unresolved because plausible is not the same as positively identified.
The viral circulation of AI-enhanced imagery adds a second layer of significance. Once generative processing is applied, apparent details can no longer be assumed to derive from the original sensor. The correct analytical object is the government source file, not the most visually impressive derivative.
The 2019 eastern United States Range Fouler case is more difficult. The witness was highly experienced and described flight characteristics outside his experience. The public video and debrief make the event worthy of further analysis. They do not provide enough range and geometry to validate a physical high-speed estimate independently. The 2025 Indo-Pacific clips similarly demonstrate unresolved military sensor events but do not publicly establish extraordinary performance. The Yellow Sea video's six-point appearance is a property of the displayed contrast pattern until optical and sensor effects are excluded. The East China Sea video provides longer tracking but insufficient public geometry for a three-dimensional reconstruction. The Pantex case is important chiefly as a documented security event.
It involved radar detection, a protective-force response and visual observation at a nuclear-weapons site. Those features strengthen the reality of the intrusion while leaving object identity open.
The Los Alamos conference is historically stronger still as documentary evidence. It proves that senior scientists and security officials met in secret to discuss unexplained green-fireball reports near a strategically critical laboratory.
It does not prove that the fireballs were engineered objects. Taken together, Release 04 demonstrates why provenance and evidential class matter more than visual drama. Some of the most sensational imagery is poorly constrained. Some of the least visually dramatic documents contain the strongest historical evidence about what government officials actually knew and did. The tranche should therefore be regarded as a substantial transparency event whose value lies in making claims more testable—not in supplying one extraordinary conclusion.
Source library
Department of War — Release 04 announcement, 10 July 2026
The controlling public announcement for the fourth tranche. It confirms the release date, identifies the release as part of PURSUE and states that additional tranches were being prepared.
Read the official Release 04 announcement
Department of War — PURSUE Release 04 archive
The primary archive for the tranche. It identifies Release 04 as cleared on 10 July 2026, provides bulk downloads of 227 MB of documents and 1.4 GB of video, and indexes records by agency, incident date, location and media type.
Browse PURSUE Release 04
Department of War — PURSUE initiative and unresolved-case definition
The controlling source for interpretation of archive status. The Department defines unresolved material as cases in which the government cannot make a definitive determination about the observed phenomenon and explicitly notes that insufficient data can be the reason.
Read the PURSUE initiative description
CBS News — fourth-release inventory and contemporary reporting
Contemporary independent reporting confirming the 40-file composition: 14 documents, 19 videos, four audio recordings and three images from five agencies. It also reports the Atlantic Range Fouler description, the 2019 experienced-aviator case, the Indo-Pacific videos and the Los Alamos document.
Read the CBS News report
AARO — UAP Report Documents
The most important government source for the modern military video descriptions. It provides AARO's neutral timelines for PR104, PR105, PR113, PR115, PR116 and other Release 04 records; documents prior digital alteration where relevant; and repeatedly warns that descriptive text is not an investigative conclusion.
Browse AARO's UAP Report Documents
DVIDS / AARO — DOW-UAP-PR116, Atlantic Ocean, 2020
Primary government-hosted copy of the 32-second Atlantic infrared video. The accompanying description records the paired Range Fouler report: maroonish appearance, estimated 12-15-foot height, wind-correlated movement, no manoeuvre or direction change and a resemblance to a large deformed balloon.
View the official PR116 video
Department of War / AARO — DOW-UAP-D091, Atlantic Range Fouler Debrief
Primary narrative record paired with PR116. Best accessed through the PURSUE archive search. The document preserves the military observer's own description and AARO's caution that descriptive estimates are not conclusive determinations of intrinsic object features.
Search the official PURSUE archive for DOW-UAP-D091
DVIDS / AARO — DOW-UAP-PR112, Eastern United States, 2019
Primary government-hosted video paired with the report from a military aviator who described the observation as unlike anything seen during 28 years of Air Force and Navy service. The source preserves the witness account without supplying a validated physical-speed calculation.
View the official PR112 video
DVIDS / AARO — DOW-UAP-PR104, Yellow Sea, 2025
Primary 18-second Indo-Pacific Command video. AARO describes an area of contrast resembling a six-pointed star, wording that appropriately distinguishes image-plane appearance from demonstrated physical morphology.
View the official PR104 video
AARO / DVIDS — DOW-UAP-PR105, East China Sea, 2025
Primary five-minute Indo-Pacific Command recording. AARO's description documents tracking, changes in zoom and display settings and intermittent loss of target distinctiveness without presenting an analytical conclusion about object identity or extraordinary performance.
View PR105 through AARO's transparency portal
Department of Energy / PURSUE — DOE-UAP-D005, Pantex Unidentified Object Incident Report, 2015
Primary government record concerning the 1 September 2015 Pantex incident. It documents radar detection, a protective-force response and visual observation at a sensitive nuclear-weapons facility. The object remained unidentified in the released record; its slow speed does not itself indicate anomalous performance.
Search the official PURSUE archive for the Pantex report
Department of Energy / PURSUE — DOE-UAP-D004, Los Alamos Conference on Aerial Phenomena, 1949
Primary historical document recording the classified February 1949 Los Alamos conference on green fireballs and related aerial phenomena. Participants included military, security and prominent scientific personnel. The document is evidence of serious official investigation, not a government finding of extraordinary craft.
Search the official PURSUE archive for the Los Alamos conference
Disclosure Archive — searchable transcription of DOE-UAP-D004
Independent searchable rendering of the released Los Alamos document, useful for navigating the historical transcript and checking participant statements against the scanned source. The PURSUE copy remains the primary record.
Read the searchable Los Alamos transcription
Department of War / PURSUE — early Air Force and Project Sign records
Release 04 contains DOW-UAP-D093, D094 and D097 covering early Air Force flying-object analysis and Project Sign, plus D092 concerning the 1966-67 Blue Book review. These are valuable primarily as historical intelligence and administrative evidence.
Browse the official Release 04 record index
Department of War / NASA — STS-80 imagery
The Release 04 portal includes three NASA images, D030-D032, described as unidentified-object imagery from Space Shuttle Columbia's 1996 STS-80 mission. Their inclusion establishes archival responsiveness to PURSUE, not an exotic physical identification.
Browse the official Release 04 archive
EarthSky — coverage of the AI-enhanced “floating brain” circulation
Contemporary astronomy-oriented reporting noting that an AI-enhanced version of the Atlantic footage went viral after the official release. Used only for the public-media context; the original DoW/AARO video and debrief remain the evidential source.
Read the EarthSky coverage