Epstein Text Messages: Michel de Cryptadamus Makes the Epstein Files Readable and Searchable
The Epstein Text Messages archive is an independent research project that transforms raw Jeffrey Epstein communications into readable, color coded, and chronologically organized webpages.
Created by the researcher and developer who publishes as Michel de Cryptadamus, the project began as an effort to reformat text message logs released by the House Committee on Oversight and Government Reform in November 2025.
It has since developed into a much larger Epstein research system. The website now organizes emails, text messages, government records, phone data, biographies, document notes, metadata, and selected files from the Justice Department’s January 2026 release.
The complete source code is publicly available through GitHub. This makes the project unusually transparent and allows other researchers to examine, reproduce, correct, or expand its document processing methods.
Snapshot
Resource name: Epstein Text Messages
Website title: The Epstein Files and Epstein Curated
Website: MichelCrypt4d4mus.github.io/epstein_text_messages
Creator: Michel de Cryptadamus
Resource type: Open source document processing and research platform
Original focus: Jeffrey Epstein’s text message logs
Current scope: Text messages, emails, phone data, government documents, biographies, timelines, metadata, document notes, and topic collections
Primary sources: The 2025 House Oversight release and a curated selection from the January 2026 Justice Department release
Source code: GitHub repository
Software license: GNU General Public License version 3
Public access: Available without registration
Research model: Automated processing combined with manual attribution, configuration, review, annotation, and correction
Important limitation: The website contains a selected and curated portion of the Epstein files, not every released document
What Is the Epstein Text Messages Archive?
The project began with a basic accessibility problem.
The text messages released by Congress appeared in machine generated logs. Those records were difficult to follow because messages were not displayed like ordinary conversations. Participants were frequently unidentified or redacted. It was also easy to confuse messages sent by Epstein with messages sent to Epstein.
Michel de Cryptadamus converted the logs into a more familiar conversational format. The reformatted pages identify the apparent participants, separate incoming and outgoing messages, organize conversations chronologically, and use color to make names and recurring subjects easier to recognize.
The original reformatted text message collection remains one of the project’s central resources.
The website later expanded into a broader system for processing and displaying Epstein related communications.
Who Is Michel de Cryptadamus?
The project is published under the name Michel de Cryptadamus. The creator also uses the social media handle Cryptadamist and publishes The Cryptocalypse Chronicles.
The public project pages link to the creator’s Substack explanation, Mastodon account, social media posts, and GitHub profile.
The creator’s publicly available work focuses heavily on cryptocurrency, technology, political influence, Steve Bannon, international power networks, and Epstein’s connections to those subjects.
The website identifies itself as a curated collection shaped by those research interests. It specifically warns that not every Epstein file is included.
EpsteinWiki could not independently verify a legal name for the person publishing as Michel de Cryptadamus. The pseudonym does not prevent researchers from evaluating the project because its code, changes, documentation, and source mapping are publicly available.
What the Website Contains
The homepage functions as a directory for several different views of the evidence.
The Most Interesting view presents a curated selection of documents chosen by the creator.
The Chronological Curated view places selected emails, text messages, and other records into a single timeline.
The Curated Emailers view groups selected communications by person.
The All Emailers view organizes the larger email collection by participant.
The Chronological Emails view places the processed email collection in date order.
The Text Messages view reformats the released message logs into readable conversations.
The Other Files view collects records that are not classified as emails or text messages.
The People view provides short biographical descriptions and links for people appearing in the archive.
The Document Notes view contains the creator’s observations about selected files.
The Phone Numbers view organizes numbers called by phones associated with Epstein.
The Communication Word Count view provides quantitative information about terms used in the communications.
The site also offers topic collections concerning money, cryptocurrency, and messages involving women and girls.
The Reformatted Text Messages
The project’s most important accomplishment is making the released text message logs readable.
The original records can contain timestamps, telephone numbers, message direction codes, account identifiers, and fragmented conversation data. Those fields are valuable to technical investigators but difficult for ordinary readers to follow.
The reformatted version converts the raw records into something resembling a normal conversation. This makes it easier to understand timing, message direction, repeated contacts, changing subjects, and conversational context.
It can also reduce one of the most serious interpretation errors in message research: attributing a statement to Epstein when it was sent to him by someone else.
The creator has openly acknowledged making that mistake during the project’s early stages. In the November 2025 explanation, Michel de Cryptadamus stated that some early screenshots confused Epstein’s messages with Steve Bannon’s messages. Publicly acknowledging and correcting an attribution error is an important sign of research transparency.
It is also a warning that every attribution should be compared with the original record.
Steve Bannon’s Prominence in the Released Messages
Michel de Cryptadamus originally reported identifying 2,108 text messages sent by people other than Epstein in the November 2025 archive.
According to that analysis, approximately 1,255 were attributed to Steve Bannon. Melanie Walker reportedly accounted for approximately 340 messages, while the remaining identified correspondents appeared much less frequently.
These figures describe the creator’s analysis of the released archive. They do not necessarily represent every message Epstein sent or received during his lifetime.
The distribution may reflect what Epstein’s estate preserved, what congressional investigators obtained, and what the House Oversight Committee selected for release.
A large number of messages demonstrates frequent communication within the released material. It does not independently prove agreement, criminal knowledge, or participation in Epstein’s offenses.
For broader context, EpsteinWiki examines Jeffrey Epstein’s political and influence networks and how Epstein’s business and trafficking system worked.
Mapping House Oversight Files to EFTA Records
The website preserves both House Oversight identifiers and EFTA identifiers when corresponding copies have been located.
This is valuable because the same underlying record may appear under different identifiers in separate government releases. A House Oversight file may later reappear as an EFTA document in a Justice Department data set.
The reformatted messages include records associated with evidence files such as:
Researchers should compare the reformatted conversation with the original evidence page before quoting a message or publishing an attribution.
Phone Records and EFTA01242527
The website includes a separate interface built around phone data in EFTA01242527.
The formatted phone number page attempts to organize numbers called by phones associated with Epstein.
Telephone records can help establish communication patterns, timing, and possible contact between individuals. However, a telephone number alone does not prove who physically used a device during a particular call.
A subscriber name does not always identify the caller. Phones can be shared, forwarded, reassigned, or used by assistants. A record of contact also does not establish the content of a conversation.
The formatted page is useful for locating patterns. The original EFTA file remains the evidence.
Chronological Research Across Multiple File Types
One of the project’s most useful features is its ability to place emails, messages, government documents, and other files into a combined chronology.
This can help researchers identify activity surrounding a specific event.
For example, a researcher studying a meeting can compare calendar messages, travel arrangements, financial transfers, email discussions, and later legal records within the same time period.
Chronological organization can reveal patterns that remain hidden when documents are separated by data set or file type.
Dates still require verification. Email chains may preserve older quoted messages. A file creation date may differ from the date of the communication. Scanning metadata may reflect when a record was processed rather than when the event occurred.
Color Highlighting and Entity Recognition
The project uses color to distinguish names, subjects, and categories.
Color highlighting can make repeated names, countries, companies, currencies, and topics visible inside long communications. This allows a researcher to scan a large record more quickly.
The software also contains data structures for identities, aliases, email addresses, telephone numbers, and recurring participants.
Entity recognition carries risks. A nickname may refer to more than one person. A person may use multiple addresses. Two people may share a surname. A name inside a quoted article is not necessarily a participant in the communication.
The website’s metadata and public code help researchers understand how an attribution was made. That transparency is one of the project’s most important strengths.
The Dataset Is Curated and Incomplete
The website prominently warns that not all Epstein files are included.
The collection contains the complete 2025 House Oversight tranche processed by the project, along with a curated selection from the Justice Department’s 2026 release.
The creator states that the selected Justice Department material reflects a particular interest in cryptocurrency and women who speak Russian.
That curatorial focus can produce valuable research. It can also shape the patterns a reader sees.
A subject may appear especially prominent because the collection deliberately includes related records. Another subject may appear unimportant because relevant files were not selected or processed.
The website should not be used to calculate the total number of communications in the complete Epstein archive unless the specific underlying dataset is clearly defined.
The Project Is Open Source
The project’s GitHub repository contains the processing code, documentation, change history, tests, deployment instructions, and a packaged research data model.
The software can generate highlighted documents, search for words or patterns, display particular files, compare duplicate records, count word usage, process message logs, and organize communications by person or date.
Technically experienced researchers can download the code and process their own copies of the source records.
The project uses the GNU General Public License version 3. This allows the code to be inspected, modified, and redistributed under the terms of that license.
Public code does not guarantee factual accuracy. It does make the project’s methods more inspectable than a closed research platform.
The Change History Shows Continued Development
The GitHub release history documents the project’s expansion.
Later versions added support for Justice Department EFTA records, chronological views, mobile output, biographical panels, document categories, email aliases, uncertain dates, uncertain recipients, phone number analysis, document notes, and additional source types.
The software includes fields for identifying uncertain dates and uncertain recipients. That is a particularly important design choice because ambiguity is unavoidable in reconstructed communications.
The change history also records fixes and methodological improvements. Researchers can use it to understand when a feature or correction entered the project.
Research Notes and Editorial Interpretation
The website includes editorial notes and curated selections created by Michel de Cryptadamus.
These notes can help identify unusual documents and explain why a record may matter. However, they represent the creator’s analysis rather than an official government finding.
The accompanying Substack articles use a highly opinionated and satirical voice. They also advance hypotheses concerning cryptocurrency, political operations, international influence, and the relationships among Epstein, Bannon, and other figures.
Those interpretations should be evaluated separately from the underlying documents.
A strong claim can still point toward genuine evidence. A persuasive narrative can also extend beyond what a document proves. EpsteinWiki recommends citing the original communication and clearly labeling any interpretation drawn from it.
Important Attribution Risks
The project improves attribution, but it cannot eliminate every uncertainty.
Redacted names may be reconstructed from context. Telephone numbers may be assigned to an apparent owner. A message direction field may be misunderstood. Similar copies of a conversation may contain different metadata.
The creator’s early confusion between Epstein and Bannon demonstrates how easily message direction can be reversed.
Before publishing a message, researchers should verify:
- The original document identifier
- The date and time
- The sender
- The recipient
- Whether the participant identification is confirmed or inferred
- Whether the message appears in a longer conversation
- Whether an attachment or earlier quoted message changes the meaning
- Whether another government release contains a clearer copy
- Whether the displayed record was manually corrected
- Whether the creator has marked the attribution or date as uncertain
How Researchers Should Use the Tool
Begin with the curated or chronological view to identify a conversation, person, event, or subject.
Open the full message chain rather than relying on a single highlighted quotation.
Record the House Oversight or EFTA identifier displayed with the conversation.
Locate the corresponding evidence file on Epstein Data.
Compare the reformatted version with the original record. Confirm message direction, participant identity, date, surrounding messages, and any redactions.
Search for the same people and dates in the site’s email, phone, and document views.
When publishing, cite the evidence file. Credit Michel de Cryptadamus when the project’s formatting, attribution work, code, or analysis materially helped locate the evidence.
Relationship to EpsteinWiki and Epstein Data
The Epstein Text Messages project, Epstein Data, and EpsteinWiki perform complementary functions.
The Epstein Text Messages archive processes communications and makes them easier to read, group, compare, and search.
Epstein Data provides direct evidence pages for EFTA records and investigative datasets.
EpsteinWiki places those records within structured articles about people, institutions, financial systems, properties, investigations, and documented events.
A researcher can use Michel de Cryptadamus’s project to discover and understand a conversation, Epstein Data to inspect the underlying evidence, and EpsteinWiki to connect the record with the wider network.
Key Takeaways
- The Epstein Text Messages archive reformats difficult message logs into readable conversations.
- Michel de Cryptadamus created the project after the November 2025 House Oversight release.
- The website has expanded to include emails, timelines, phone records, biographies, metadata, document notes, and selected Justice Department files.
- The archive combines automated processing with manual attribution, configuration, review, and correction.
- The project is open source, allowing researchers to inspect and reproduce its methods.
- The website clearly warns that it does not contain every Epstein file.
- The Justice Department portion is curated around the creator’s interests, particularly cryptocurrency and Russian speaking women.
- The creator publicly acknowledged early errors involving the direction and attribution of messages between Epstein and Steve Bannon.
- Reformatted messages are research aids. Original House Oversight and EFTA records remain the evidence.
- The project is especially valuable for chronological research, participant comparison, message direction, and connecting duplicate files across government releases.
Why This Project Matters
Government transparency is not meaningful when public records are technically available but nearly impossible to read.
The original text message logs required readers to decode metadata, determine message direction, identify participants, and reconstruct conversations manually. Michel de Cryptadamus converted that material into a form ordinary researchers can follow.
The project does more than improve appearance. Readability affects accuracy. A clear conversation makes it easier to recognize context, sarcasm, changes in subject, and the difference between a sender and a recipient.
Its greatest strength is not that it tells researchers what the records mean. Its greatest strength is that it makes the records easier to examine for themselves.
Sources
- Epstein Text Messages and Epstein Curated homepage
- Reformatted Epstein text messages
- Chronological curated records
- Michel de Cryptadamus project explanation
- Epstein Text Messages GitHub repository
- Project release history
- Epstein Data EFTA01242527
- Epstein Data EFTA01218267
- Epstein Data EFTA01214317
- Epstein Data EFTA01209003
- Epstein Data EFTA00783435
- Epstein Data EFTA00786405
- Epstein Data EFTA00507900
- EpsteinWiki Knowledge Base