89,901 Pages: What's Actually in the Stargate Documents

When the CIA released the CREST database in January 2017, the resulting coverage mostly followed one of two tracks: either it was a story about government credulity — tax money spent on psychics — or it was a story about suppressed proof of paranormal phenomena. Both framings treat 12,473 separate documents as a single thing with a single meaning. They don't.

The archive is a collection of records produced over 23 years by multiple agencies, multiple contractors, multiple research programs operating under different names and different methodologies. Reading it as a unified document misses what makes it interesting: the internal disagreements, the shifting protocols, the moments where the researchers themselves were not sure what they were looking at.

The FOIA Release and John Greenewald's Role

The Stargate archive exists in its current accessible form because of a single person's sustained effort. John Greenewald Jr. began filing Freedom of Information Act requests at age 15 in 1996, eventually creating The Black Vault — what would become the largest online FOIA document clearinghouse in the world. For years, Greenewald fought the CIA's resistance to releasing the full scope of Stargate documentation. In January 2017, after a years-long FOIA effort, the agency compiled its complete research material from the remote viewing program into a multi-disc release: 89,901 pages spanning 12,473 separate documents.

Rather than stopping at the government's native format — the CIA's CREST (CIA Records Search Tool) database delivers the documents as .tif image files, which are difficult to search and require specialized viewing software — Greenewald undertook the labor of converting all 89,901 pages into searchable PDFs. This version at theblackvault.com remains the most accessible path into sustained engagement with the archive. The Internet Archive also hosts the full browsable dataset at archive.org/details/STARGATEDataset, making the collection available in perpetuity.

How to Navigate the Archive: Access Points and Limitations

Three main interfaces exist for accessing the Stargate documents, each with different strengths and weaknesses. The CIA's own reading room at cia.gov/readingroom/collection/stargate contains the authoritative 12,473 source documents. Its advanced search feature allows filtering by date range and document type, but the native .tif format creates a bottleneck for serious research. The CREST (CIA Records Search Tool) database itself predates the Stargate release and contains broader declassified CIA records; the Stargate collection is a subset, though the CREST interface suffers from the same full-text search limitations that plagued the original design.

Greenewald's Black Vault searchable PDFs solve the full-text search problem, making it possible to search for specific names, locations, methodology terms, or target coordinates across the entire collection. For researchers without the time to download and index the full dataset, the Black Vault version represents a practical middle path between official completeness and genuine usability.

The Program Name Chronology: Six Decades, Six Names

To navigate the archive's 23-year span intelligently, you need to understand the institutional history behind the code names. The program existed under six different designations, each corresponding to shifts in funding agency, sponsorship, and operational focus.

SCANATE (1972–1975): The program began in 1972 as a CIA initiative called SCANATE, short for "SCANning with coordinATEs." Researchers at Stanford Research Institute (SRI International) in Menlo Park, California started the fundamental research into whether remote viewing could be done reliably, and whether it could be taught. In this phase, the focus was exploratory and scientific — no operational tasking, just the basic R&D question: is this real, and if so, how does it work?

GONDOLA WISH (1977): By the mid-1970s, the Army's Assistant Chief of Staff for Intelligence (ACSI) was watching the SRI results with interest. In 1977, they launched GONDOLA WISH as a preliminary evaluation effort to assess whether remote viewing could have operational applications in military intelligence. This was a bridge phase — still exploratory, but with military stakeholders in the conversation for the first time.

GRILL FLAME (1978–1983): Building on GONDOLA WISH's findings, GRILL FLAME became the first operational collection project. Formalized in mid-1978 under Army intelligence command (INSCOM), it integrated the SRI research program into a structured, tasked operation. The Army deployed the first operational remote viewing unit at Fort Meade, Maryland. Hundreds of remote viewing experiments were carried out at SRI through 1986, with the results flowing into an intelligence pipeline for the first time.

CENTER LANE (1983–1985): In 1983, the Army redesignated the program as CENTER LANE under INSCOM (ICLP — INSCOM CENTER LANE Project). The transition was mostly administrative, but it marked the consolidation of military interest and a shift toward treating remote viewing as an established operational tool rather than an experimental research project.

SUN STREAK (1985–1991): In late 1985, Army funding was terminated, but the program did not die. Instead, it was redesignated SUN STREAK and transferred to the Defense Intelligence Agency (DIA), specifically to the Scientific and Technical Intelligence Directorate. This phase represents a crucial transition: from a research program trying to prove the concept, to an operational program trying to build a scalable capability. The DIA attempted to train new remote viewers from scratch, developing structured training protocols and testing the extent to which CRV could be taught to military personnel without pre-existing psychic abilities.

STARGATE (1991–1995): The final phase began in 1991 when the program was transferred to Science Applications International Corporation (SAIC), a major defense contractor. It was renamed STARGATE and operated under DIA sponsorship through 1995, when Congress and the CIA commissioned a final review. The unclassified summary of that review concluded the program had no documented intelligence value, and it was terminated.

What the Documents Actually Contain

The 89,901 pages break down into several document categories, each revealing different layers of the program's reality.

Session Transcripts: These are the raw output of remote viewing operations. A typical transcript records a viewer sitting in a room with a blind target (coordinate-based, or a sealed envelope, depending on the protocol being tested), and describes impressions that arise. The transcript captures the viewer's verbal output, the monitor's observations, and often includes sketches. The coordinate format used in many later sessions looks like arbitrary six-digit numbers — the actual geographic coordinates of the target location, kept hidden from the viewer. Some transcripts are a few pages; others run to 20 or 30 pages as the viewer works through progressively refined impressions.

Quarterly Reports to Congress: Among the archive's most revealing documents are the briefing papers prepared for congressional oversight committees. These show how program managers explained their work to elected officials who controlled funding. The reports are relatively honest about the variation in results, the difficulty of operational tasking, and the tension between research findings and operational expectations. They also reveal the bureaucratic reality: meetings scheduled, budget cycles, personnel turnover, and the constant negotiation between what the science could support and what the military wanted to believe.

Budget Documents: The archive includes year-by-year budget summaries showing that the entire 23-year program cost approximately $20 million — a relatively modest sum in intelligence terms. The budgets reveal staffing levels (at its peak, around 23 remote viewers and 40+ total personnel), equipment costs, and contractor payments to SRI and later SAIC. These figures help ground the program's actual scope, which was smaller and more understaffed than popular accounts often suggest.

Internal Evaluations and Methodology Reports: The archive contains dozens of internal reviews where program managers, contractors, and external evaluators assessed the program's effectiveness. These documents reveal a pattern of genuine uncertainty. Some evaluators reported significant statistical anomalies in the data; others raised methodological concerns. There are memos from skeptical officers questioning whether the resources were well spent, and detailed technical reports from SRI physicists breaking down the CRV protocol stage by stage.

Correspondence Between Sponsors and Researchers: Letters and memos between CIA/DIA program officers and the research teams at SRI show the tension between research integrity and operational pressure. Researchers sometimes resisted tasking they felt was inappropriate; program managers sometimes pushed for results they wanted regardless of the data; sponsors asked for evidence of strategic value to justify continued funding.

The Coordinate Remote Viewing Protocol: How It Worked

To understand what the session transcripts are recording, you need to know the structure of Coordinate Remote Viewing. Ingo Swann, working with physicist Hal Puthoff at SRI, developed CRV in the early 1980s as a response to a military directive: develop a teachable, reliable method for remote viewing. The protocol was not mystical; it was intended to be as systematic and learnable as any intelligence tradecraft.

CRV consists of six sequential stages. Stage 1 focuses on the ideogram — a spontaneous doodle or sketch the viewer produces in response to the coordinate, capturing the "flavor" of the target location without visual detail. Stages 2 through 6 progressively refine and elaborate on that initial impression, moving from sensory data (colors, textures, temperatures) through dimensional information (sizes, distances, angles) to conceptual and analytical overlay (interpretation of what is being perceived). The method explicitly acknowledges that the human mind will try to interpret and "fill in" missing information — what Swann called AOLs (Analytic Overlays) — and the protocol is designed to separate genuine perceptual impressions from analytical guessing.

From 1982 to 1984, five military personnel from Fort Meade underwent formal CRV training contracted with SRI International. Later, during the SUN STREAK and STARGATE phases, the DIA attempted to train dozens of new viewers using CRV. The archive shows exactly what happened: some trainees developed genuine facility with the protocol; others showed little progress. The documents suggest that individual aptitude mattered, but that CRV also democratized the skill somewhat — it was teachable in a way that earlier freestyle remote viewing was not.

Notable Sessions and Documents Worth Finding

Several sessions in the archive have attracted attention from independent researchers and deserve particular scrutiny.

The 1973 Jupiter Session: In April 1973, Ingo Swann performed a remote viewing session on the planet Jupiter, years before the Voyager probe provided verification. In a 20-minute session, Swann outlined thirteen atmospheric and structural features, many of them speculative or contradicting the prevailing scientific consensus at the time. Later, NASA's Pioneer and Voyager missions confirmed several of his observations, including Jovian rings — which the scientific establishment in 1973 considered impossible. The session is available in the Black Vault collection and is striking both for the specificity of the descriptions and for the epistemological problem it raises: how did he know that?

The Semipalatinsk Sessions: Among the operational tasking sessions, remote viewer Pat Price's work on the Soviet weapons facility at Semipalatinsk in the mid-1970s became legendary in RV literature. Over three days, Price sketched and described a mobile gantry crane at the secret facility with engineering-drawing accuracy — 150 feet tall with railed tracks 50 feet apart. An evaluating officer noted that Price's sketch seemed "inconceivable" without either actual remote viewing or a security breach. The incident triggered a formal Congressional investigation to determine if classified information had been leaked; none was ever found. These sessions, though now over 50 years old, remain the closest thing to "operational proof" in the public archive.

1984 SRI Training Reports: The SRI methodology reports from 1984, particularly the "Special Orientation Techniques" document, show the development of the ideogram protocol through over 140 experimental sessions. The archive reveals that researchers worked through more than 1,000 sessions to refine and validate the CRV structure. These documents are valuable for understanding that CRV did not emerge fully formed; it was iterated, tested, and debugged like any research protocol.

The DIA Phase: Operationalizing Remote Viewing

The transition from SRI research to DIA operations reveals both the program's strength and its central tension. SRI had demonstrated that remote viewing was statistically anomalous. The military and intelligence community wanted to know whether it could be weaponized — whether a scalable cadre of trained viewers could be deployed as an intelligence asset.

The SUN STREAK and STARGATE phases attempted exactly that. The DIA recruited candidates, trained them in CRV, and tasked them with operational intelligence questions: What is the layout of this facility? What weapons are in this location? What happened at this event? The archive shows that the attempt met with mixed results. Some trained viewers showed facility with the protocol and produced assessments that evaluators found useful. Many did not; the training produced limited results for some candidates regardless of effort invested. The program peaked during the mid-1980s with approximately seven full-time remote viewers and 40+ total personnel, making it a modest but real operational program.

By the early 1990s, skepticism was increasing. The CIA commissioned the Cognitive Sciences Laboratory at SAIC to conduct the final programmatic review. The 1995 unclassified summary concluded that while remote viewing produced statistically anomalous results in laboratory conditions, "there is no evidence that the program has ever provided actionable intelligence to intelligence community customers." That assessment triggered the program's termination, though the declassification occurred later.

Internal Disagreements and Skepticism in the Archive

One of the archive's most honest features is its record of internal disagreement. The program did not exist in a bubble of true believers. Physicists, intelligence officers, and evaluation personnel raised rigorous skeptical objections throughout.

Program managers disagreed on methodology. Some officers questioned whether the training time and resource investment could be justified. External reviewers — including a 1995 National Academy of Sciences review — critiqued the lack of rigorous scientific validation and argued that "successes" could be attributed to statistical chance, suggestive questioning during debriefing, or retroactive fitting of results to actual events.

The archive also documents a significant structural weakness: program principal investigator Edwin May served simultaneously as the primary researcher, the judge of data quality, and the gatekeeper for independent evaluation. This conflict of interest — May presided over 70% of contractor budget and 85% of data collection — troubled external reviewers. One independent evaluator, David Marks, noted that May refused to release the names of the oversight committee and denied permission for independent judging of transcripts, commenting that "this refusal suggests that something must be wrong with the data or with the methods of data selection." That internal contradiction is now part of the permanent record.

What's Notably Absent: Redactions and Classified Gaps

The archive is not perfectly transparent. Operational targeting documents — the actual intelligence tasking orders and real-world applications of remote viewing — are mostly still classified. Text is occasionally redacted in the middle of documents, with annotations (if decipherable) offering clues to the reason: protecting operational details, shielding human intelligence sources, concealing ongoing methods.

One of the frustrations for researchers is that the actual target of many sessions is missing from the released documents. A transcript might describe impressions without clearly stating what the viewer was supposed to be describing. However, quarterly reports and evaluations often provide context clues, and the archive's sheer size means that cross-referencing across documents can sometimes reconstruct the picture.

What is available has been released with relative openness. According to one official involved in the declassification process, the instructions were to protect human sources while releasing the bulk of the research and operational material. Compared to other classified intelligence programs, the Stargate archive represents an unusually transparent window into a sensitive program's actual methods, results, and internal debates.

Reading the Archive with Open Eyes

The most honest thing you can say about the Stargate archive is that it documents something that resisted simple explanation for 23 years while being closely observed by physicists, military intelligence officers, congressional oversight staff, and multiple independent evaluators. Statistical anomalies appeared in the data. Some operational sessions produced details that seemed difficult to explain through conventional means. Other sessions generated only noise. Training showed it could be taught, but not everyone could learn it. Congress cut funding anyway.

The archive does not prove remote viewing works. The 1995 review was correct: no actionable intelligence is documented as having flowed from the program's results. But the archive also does not prove remote viewing is impossible. What it shows, across 89,901 pages, is that a serious government program spent $20 million and 23 years investigating something genuinely anomalous, and found that the anomaly was real but operationally unreliable.

Anyone who tells you the archive definitively proves or definitively disproves remote viewing has not actually read it. The honest position, after reading it, is to acknowledge that the question remains open — but that the evidence available to us is, at minimum, worth taking seriously.