"I Deleted Evernote After 12 Years" - Migration Stories from VaultBook Users

The decision to leave a note-taking application you have used for years is rarely made in a day. It usually arrives after a long accumulation of small discomforts - a terms of service update that you read more carefully than usual, a news story about the platform that surfaces a concern you had always half-suppressed, a moment where you search for something important and realize it is buried somewhere in a cloud you do not control. The discomfort builds until the calculus shifts: the friction of leaving feels smaller than the friction of staying.

The people in this article reached that calculus at different points and for different reasons. Marcus Chen spent twelve years building a legal reference library in Evernote. Priya Raghavan spent seven years managing her professional life in Notion. David Harrington spent eight years documenting investigations in Microsoft OneNote. Elena Vasquez spent five years keeping clinical notes in Bear with iCloud sync. James Okafor spent nine years accumulating academic research in Google Keep and Google Docs. All of them now keep their notes in VaultBook, on their local devices, with no cloud sync and no server receiving any of their content.

These are their stories - told in their own words, in enough detail to be useful to anyone standing at the same decision point they once faced.

Marcus Chen: Immigration Attorney, 12 Years in Evernote

Marcus Chen has been practicing immigration law for nineteen years. For twelve of those years, he kept his professional reference library in Evernote: annotated regulatory guidance, USCIS policy memoranda, historical visa bulletin archives, country condition reports for asylum cases, procedural notes from filing officers he had worked with across dozens of immigration courts, and the accumulated institutional knowledge of a career spent navigating one of the most complex and frequently changing regulatory environments in American law.

“I was not a casual Evernote user,” he says. “I had over 4,000 notes organized into a notebook structure I had refined across a decade. I had a search-first workflow. I knew exactly how to find things. It worked.”

The first crack appeared in 2023 when Evernote’s parent company changed its terms of service in ways that made Marcus uncomfortable. “I read the updated terms more carefully than I had read the original ones. There was language about using content to improve their services that I had not thought carefully about when I signed up. I started asking myself: what exactly are they doing with 4,000 notes about immigration law, asylum country conditions, and individual client procedures? I did not have a good answer.”

The crack widened when Evernote’s service reliability deteriorated. Notes failed to sync across devices. The search index occasionally produced results that were a day or two behind the actual state of his notes. For a lawyer whose practice depends on being able to retrieve a specific policy memorandum at the moment a filing officer raises a question, that lag was professionally significant.

The breaking point was a conversation with a client whose asylum case involved country condition evidence Marcus had documented extensively over several years of case work. He could not find a specific note during a client call, searched repeatedly, and ultimately had to call back after finding it manually. The note was in Evernote. It had always been in Evernote. He just could not locate it when he needed it.

“That call was the moment. I had spent twelve years building a research library and I was not able to retrieve a document during a live client conversation. The tool had failed the one job I needed it to do.”

The migration took four weeks. Marcus exported his entire Evernote library as ENEX files, converted them to Markdown, and imported the Markdown files through VaultBook’s Import from Obsidian tool, which processes standard .md files regardless of their origin. The initial import - approximately 3,800 notes that survived the format conversion cleanly - took one session. The remaining 200 notes with complex formatting required manual recreation, which he did incrementally over the following weeks.

He rebuilt his organizational structure using VaultBook’s nested page hierarchy. The notebook system he had developed in Evernote over twelve years mapped cleanly onto the page architecture: each major practice area became a top-level page, each sub-area became a nested child page, and the drag-and-drop reordering allowed him to arrange the pages in the order that matched how he actually navigated them during work. The color coding and icons made the visual navigation faster than Evernote’s notebook list had ever been.

The search immediately impressed him. VaultBook’s relevance scoring - title matches weighted at 8, labels at 4, body text at 3, attachment text at 1 - surfaces the most relevant results first rather than returning a flat chronological list of matching notes. The typeahead dropdown begins suggesting results as he types, often surfacing the right note before he finishes the query. “The search is better than what I left. Noticeably better. I can find a policy memorandum from 2018 using two words from its title and it is the first result.”

For the most sensitive case materials - documents that contain client-adjacent information or that are relevant to asylum cases where the safety of the client’s family in their home country depends on confidentiality - Marcus uses VaultBook’s per-entry AES-256-GCM encryption. Each sensitive entry carries its own password, distinct from any other entry. The session password caching means he enters the password once per session and can then open and edit the entry freely until the browser closes.

The version history, available for every entry with a 60-day retention window, has proven valuable in ways he did not anticipate. Immigration law changes frequently - a policy memorandum gets superseded, a court decision shifts the interpretation of a regulatory provision, a country condition changes between the time a case is filed and the time it is decided. When Marcus updates a reference entry to reflect new guidance, the version history preserves the previous version of the entry, which matters for documenting what the applicable guidance was at the time a specific filing was made.

The labels system has replaced what Evernote called tags, with considerably more organizational power than the flat tagging he had used before. His label vocabulary is now a parallel classification layer running on top of the page hierarchy: cite-ready for entries whose content has been verified and is ready to use in client communications, needs-update for entries that contain regulatory guidance that may have changed, procedural for entries that document court or agency procedures rather than legal standards, and country-condition for the asylum-related research entries. The label match mode - filtering for entries carrying all of a specified set of labels simultaneously - lets him construct compound views that the page hierarchy alone cannot provide. A filtered view of every entry labeled both country-condition and needs-update is an instant review list for entries whose research may have been overtaken by changing political conditions.

The 14-day activity chart in the Pro analytics panel shows him which research areas have had recent activity and which have gone quiet - a quiet area in the chart often signals a research thread that needs attention before a filing deadline approaches.

“I deleted Evernote three weeks after the migration was complete,” he says. “I kept it running in parallel for three weeks to make sure I had not missed anything. Then I deleted it. I have not thought about it since.”

Priya Raghavan: Product Manager, 7 Years in Notion

Priya Raghavan runs product for a mid-sized health technology company that builds workflow tools for clinical teams. She used Notion for seven years - first personally, then professionallyupon it spreading through her team until it became the unofficial operating system for their entire product management function.

The trigger for Priya was not a single event but a slow accumulation of professional risk awareness. As a product manager in health technology, she spends her days working with clinical workflows, patient data handling standards, and the regulatory requirements that govern how health-adjacent software is built and operated. She reviews HIPAA compliance documentation as a routine part of her job. And over time, she became increasingly aware that Notion’s architecture - a cloud-first workspace where all content lives on Notion’s servers, processed and indexed by Notion’s infrastructure - was incompatible with the sensitivity of what she was putting into it.

“I was storing meeting notes from conversations with clinical customers that included references to their patient workflow processes. I was storing competitive analysis that included confidential information from partner conversations. I was storing the strategic roadmap for a health technology product that is subject to regulatory oversight. None of that should be in a cloud workspace that I do not control.”

The moment that moved her from discomfort to action was reading Notion’s AI product terms carefully after the company introduced AI features into their workspace. The terms for AI-assisted features included language about content being used to train and improve AI models that she found unacceptable for professionally sensitive material. She was not alone - several of her colleagues raised the same concern at the same time - and the team began evaluating alternatives.

Her migration from Notion involved exporting her personal workspace as Markdown and importing through VaultBook’s Import from Obsidian tool. The seven years of personal notes - meeting templates, project frameworks, personal knowledge management, reading notes, and the professional reference material she had built up over her career - came in cleanly. The team’s shared Notion content, which needed to remain accessible to colleagues, stayed in Notion; her personal working vault moved entirely to VaultBook.

The adjustment that took the most time was rebuilding her organizational habit around VaultBook’s pages and labels system rather than Notion’s database-and-property model. Notion’s power comes partly from its structured database views, which allow filtering and sorting on custom properties. VaultBook’s organizational model is different: the pages hierarchy provides the topical structure, labels provide the cross-cutting classification, and the sort and filter controls provide the viewing flexibility.

“The labels system is more versatile than I initially gave it credit for,” she says. “In Notion I was maintaining database properties for everything. In VaultBook I apply labels and filter by them, and the label match mode - filtering for entries that carry all of a set of labels simultaneously - gives me the compound filter views I relied on in Notion. It is a different mental model but it gets me to the same places.”

The features that have most changed her workflow are the ones she did not have in Notion at all. The Q&A panel - the natural-language search that accepts questions and returns scored answers - has become her primary retrieval interface for the reference material that accumulates over a product management career. Asking “what did we decide about the notification architecture in the v3 redesign” returns a ranked list of entries scored on relevance, with the entries that contain the specific decision near the top. The vote-based reranking, which allows her to train the search results over time by upvoting the right answers and downvoting the wrong ones, has made the Q&A increasingly accurate for the specific vocabulary of her product domain.

The AI Suggestions carousel has become part of her morning routine. The first page of suggestions - entries with upcoming due dates and the entries the weekday pattern identifies as likely relevant today - gives her a morning briefing that she did not have to configure. “I open VaultBook in the morning and the suggestions tell me what I was working on last Tuesday, which is usually what I should be working on this Tuesday. It has made my weekly review almost automatic.”

The Kanban Board, which reads from entry labels and organizes them into drag-and-drop columns, replaced the project tracking she had been maintaining manually in a Notion database. Her board columns correspond to her workflow phases: Exploring, In Progress, Blocked, Review, Done. Changing an entry’s label on the note itself moves its card on the board automatically.

The multi-tab view has become central to how Priya manages simultaneous projects. She keeps four or five tabs open at any given time - one per active project area - each with its own independent view state, filter set, and sort order. Switching between projects is a click rather than a navigation sequence. Notion’s multi-page navigation required her to maintain browser tabs as a substitute; VaultBook’s native tab strip does the job directly.

The advanced date filters and file-type filters sit alongside the label filters in her toolbar. Filtering to entries modified in the last 7 days, carrying the In Progress label, and containing attachments gives her an instant list of active project materials that have seen recent work. She runs this filter at the start of her Friday review to see what has been touched during the week.

The Password Generator tool, tucked in the Pro Tools panel, has become her go-to for generating strong credentials for the test accounts she creates during product evaluation work. It saves a context switch to a separate password manager interface during workflows that already live in VaultBook.

“I stayed in Notion seven years because the switching cost felt high. The switching cost turned out to be about two weeks of migration work. The day I finished the migration I felt something I did not expect: relief.”

David Harrington: Investigative Journalist, 8 Years in Microsoft OneNote

David Harrington has spent the last fourteen years as a freelance investigative journalist contributing to publications that cover corporate accountability, financial fraud, and regulatory failures. For eight of those years he kept his investigation notes in Microsoft OneNote, synchronized across devices through OneDrive.

His reason for leaving was direct: source protection.

“I was doing an investigation that involved a source inside a financial institution. The source had taken significant personal risk to provide information about practices that were harmful to their company’s customers. I was keeping my notes about that source in OneNote. And OneNote syncs everything to Microsoft’s servers. Microsoft has responded to legal process from governments and corporations in the past. I was putting a source’s safety in a file cabinet that could be opened by a sufficiently motivated subpoena.”

The concern crystallized during a conversation with a lawyer who specializes in press freedom. The lawyer’s assessment was not reassuring: cloud-synced notes are subject to legal process directed at the cloud provider, in the cloud provider’s jurisdiction, and the journalist may not be notified before the process is served. The practical standard for protecting sensitive source materials is local storage that does not pass through any third-party server.

David migrated from OneNote to VaultBook in a single intensive weekend. OneNote exports as .one files, which required conversion to a text format before the Import from Obsidian tool could process them. He used a conversion utility to extract text from the OneNote files, cleaned up the output, and imported it as Markdown. The structural fidelity was not perfect - OneNote’s free-form canvas layout does not map cleanly onto linear note content - but the text content of eight years of investigation notes transferred completely.

His vault structure organizes investigations by phase. A top-level Active Investigations page contains sub-pages for each current investigation, each with its own internal structure of Pages, Sources, Documents, and Analysis. A top-level Archive page contains the same structure for investigations that have concluded or been filed. A top-level Reference page holds the research that is not investigation-specific: regulatory frameworks, corporate structure research on industries he covers frequently, past coverage of recurring subjects.

For every entry that contains source-adjacent information - which in David’s case means any note that could identify a source, characterize a source’s access level, or indicate the nature of information a source provided - he applies per-entry encryption with a unique password. The encrypted entries store only ciphertext in the vault files on his device. Without the entry password, the content is opaque.

“I have about sixty encrypted entries across my active investigations,” he says. “I enter the passwords at the start of my work day and they are cached for the session. I can open and edit those entries freely during the session. When I close the browser, the cache is gone. It is a good workflow.”

The deep attachment indexing has transformed how he manages investigation documents. PDFs of regulatory filings, court documents, corporate earnings calls, and legal decisions are attached to the relevant entries and indexed through VaultBook’s full-text extraction. Every page of every attached PDF is searchable from the main search bar. A search for a specific individual’s name reaches across every attached document in the vault, surfacing every filing, decision, or corporate document that names that person.

The OCR indexing of screenshots has been particularly valuable. Investigative journalism involves a constant accumulation of visual evidence: screenshots of online communications, captures of now-deleted social media posts, photographs of physical documents, screenshots of corporate websites before content was changed. In OneNote, these screenshots existed but were invisible to search. In VaultBook, Tesseract runs on every inline image and every attached image, and the recognized text joins the searchable index. A screenshot of a corporate website from 2022 that was captured before the page was taken down is now findable by the text that was on that page.

The Related Entries panel, which scores the entire vault against the currently open entry and surfaces the most similar notes, has caught connections between investigations that David did not notice during the work itself. A detail in a 2022 investigation that reappears in a 2025 investigation, connecting the same intermediary across different cases, was surfaced by the Related panel when he was reviewing a current entry.

“The tool that I have now is better for the work than OneNote was, independent of the privacy architecture. The privacy architecture makes it the only tool I would trust for this work. Those two things together make it the obvious choice.”

Elena Vasquez: Clinical Social Worker, 5 Years in Bear and iCloud

Elena Vasquez has maintained a private practice as a licensed clinical social worker for eleven years. She sees clients dealing with trauma, anxiety disorders, relationship difficulties, and adjustment to major life transitions. Her clinical notes contain the most sensitive category of personal information that exists: the private disclosures of people in therapeutic relationships, made under an expectation of strict confidentiality that is both ethically required and legally mandated.

For five years, she kept those notes in Bear, synced to iCloud and accessible across her iPhone, iPad, and MacBook. Bear is a well-designed, capable application with a clean interface that clinicians often favor for its simplicity. The iCloud sync meant her notes were always up to date across her devices.

The problem was that iCloud sync means the notes are on Apple’s servers.

“I had told myself a story about iCloud being encrypted and Apple being a privacy-focused company. That story was not wrong exactly, but it was incomplete. Even with end-to-end encryption on iCloud, my clinical notes were passing through Apple’s infrastructure, stored on Apple’s servers, subject to Apple’s terms of service, and accessible through Apple’s account recovery mechanisms if my account were ever compromised. HIPAA requires that I protect client health information. I was not protecting it as well as I was required to.”

The migration from Bear took one afternoon. Bear exports notes as individual Markdown files, which the Import from Obsidian tool processes directly. Elena’s five years of clinical notes - approximately 800 entries - imported in a single session, with all the note content, headings, and formatting intact.

She organized her vault into a structure that maps to the clinical workflow: an Active Clients page with a sub-page per client (named only with a client code, never a real name in any title or page label), a Closed Cases archive, a Clinical Reference page for protocols and frameworks, a Supervision Notes page for her regular sessions with her clinical supervisor, and a Practice Administration page for the operational side of running a private practice.

Every single entry in the Active Clients section carries per-entry encryption with a unique password per client. The therapeutic content of clinical notes must be protected at a level that makes unauthorized access computationally infeasible, not merely policy-restricted. AES-256-GCM with PBKDF2 key derivation at 100,000 iterations provides that protection. The key is derived from a strong password that exists only in Elena’s memory and in her encrypted password manager - it is never stored in any file on any server.

The session caching keeps the workflow practical. Elena opens VaultBook at the start of her clinical day, enters the passwords for the clients she has sessions with that day, and works through her session notes with those entries fully accessible. The lock screen - a full-page blur overlay that can be triggered instantly - is mapped to a keyboard shortcut she activates any time she steps away from her desk between sessions.

“The experience of using VaultBook is that I am doing clinical work. The security is invisible during the workflow. I enter passwords in the morning. I lock the screen between sessions. I close the browser at the end of the day. The clinical note-taking itself is just note-taking.”

The sections architecture within each client entry has allowed her to build a structured clinical record within a single note rather than maintaining separate notes for different aspects of the clinical work. Each client entry has sections for: the initial assessment and presenting concerns, session notes (chronologically organized, newest at top), treatment goals and progress tracking, significant disclosures, and her own clinical formulation notes that are maintained separately from the session record.

The version history has practical value in clinical documentation. Clinical notes are sometimes updated after a session to add context or correct an error. When an update is significant - a correction to an observation, an addition of information disclosed at the end of a session - the version history preserves both the original note and the amended version, creating a defensible documentation record that shows what was recorded when and what changes were subsequently made.

She runs the Folder Analyzer periodically to understand the storage footprint of her vault and the PDF Compress tool to reduce the file size of scanned documents before attaching them, keeping the vault folder manageable over years of accumulation.

The Timetable integrates her session schedule with the vault directly. Client sessions are created as entries with due dates matching the session time. The Timetable’s day and week views show her the day’s clinical schedule alongside the other scheduled items in her practice calendar. The Timetable ticker in the sidebar shows the next upcoming appointment at a glance, so she can see how much time remains between sessions without switching to a calendar application.

The Threads tool has replaced the legal pad she used to keep beside her keyboard for capturing thoughts during sessions that did not fit the structured clinical note format - a sudden observation about a pattern, a reference she wanted to look up, a therapeutic intervention idea to consider for a future session. The Threads overlay opens as a centered scratchpad without navigating away from whatever she was last looking at, and the content persists in the vault’s attachments/ folder as its primary storage. At the end of her clinical day, she reviews the Threads content and moves anything that belongs in a specific client entry to that entry, then clears the thread.

The AI Suggestions panel’s weekday pattern has been a quiet efficiency gain. The entries it surfaces on Tuesday mornings - the pattern built from four weeks of Tuesday session behavior - are consistently the entries she would have navigated to manually. The supervision entries she reviews before her Thursday supervision call appear in the suggestions on Thursday mornings. “It knows my week better than I do,” she says.

“What I have now is clinical documentation that is genuinely secure, organized in a way that serves the clinical workflow, and that runs entirely on my device. No Apple server. No app company server. No cloud. The notes my clients entrust me with stay on my device. That is what they deserve and what the ethical standards require.”

James Okafor: Associate Professor, 9 Years in Google Keep and Google Docs

James Okafor is an associate professor of political science at a research university, specializing in comparative democratization and authoritarian resilience. Over nine years of active research, his notes accumulated across two Google products simultaneously: Google Keep for quick captures and reference snippets, and Google Docs for longer analytical writing and literature notes. The combination worked, but barely - the two products did not talk to each other, search operated independently in each, and the organizational logic that worked in Keep was incompatible with the document structure in Docs.

“I had built a system that required me to remember which of two completely separate applications held the thing I was looking for,” he says. “Google Keep for snippets and quotes. Google Docs for analysis. Except sometimes I wrote analysis in Keep because I was on my phone, and sometimes I put snippets in Docs because I was already there. The system had a logic that I violated constantly, and then I had to search both places for everything.”

The privacy concern had always been present but suppressed. James studies authoritarian governments for a living. He corresponds with scholars, journalists, and civil society actors in countries where those activities carry real risk. His research notes contain the names of people in vulnerable positions, characterizations of political situations that could be sensitive in certain contexts, and analytical material that reflects the professional judgments of people who trusted him with their perspectives.

“I finally did a clear-eyed assessment of what I was storing in Google products. Google’s business is knowing things about people. Everything I put into Google Keep and Google Docs is available to Google’s infrastructure - for product improvement, for training AI models, and potentially for response to government legal process. For most of what I write that is fine. For some of what I write, it is not fine. I could not cleanly separate the categories in my Google workflow.”

The migration required more work than most. Google Keep does not export to a standard format without a workaround through Google Takeout, which produces a JSON archive. James wrote a short Python script to convert the JSON export to Markdown files, then imported those through VaultBook’s Import from Obsidian tool. His Google Docs - several hundred documents of varying length - he converted to Markdown manually for the most important ones and exported as text for the remainder.

The resulting vault consolidated, for the first time in nine years, all of his research material in a single searchable system. The organizational structure he built around VaultBook’s nested pages finally gave him the unified architecture he had been attempting to approximate with the two-product system: research projects as top-level pages, sub-pages for literature reviews, field notes, analytical drafts, and correspondence notes, with a cross-cutting labels system for status (active, published, dormant), type (primary-source, secondary-source, analysis, field-note), and project association.

The feature that most immediately changed his experience was the deep attachment indexing applied to his PDF library. James has accumulated thousands of academic papers in PDF format over fifteen years of research. These are his primary sources, the empirical and theoretical foundation of everything he writes. In his previous workflow, PDFs existed outside the note-taking system entirely - in a folder hierarchy or in a reference manager, searchable only by metadata and title, their full text accessible only by opening each file individually.

In VaultBook, every PDF he attaches to a relevant entry is fully text-indexed through pdfjs extraction. A search for a specific theoretical concept, a scholar’s name, or a country’s political situation reaches inside every indexed PDF and returns results from their full text. A paper he read in 2017 and attached to a literature note that year surfaces in a 2025 search for a concept the paper addressed.

For scanned PDFs - older scholarship, working papers produced before digital-first publishing, photocopied documents from archival research - the OCR pipeline renders each page at 2x scale and runs Tesseract recognition. The character recognition on aged documents is imperfect, but the resulting text index is almost always sufficient to make the document findable by its key concepts.

The Q&A panel has become central to his literature review process. When beginning work on a new paper, he uses the Q&A to ask structured questions against his vault: “which cases exemplify competitive authoritarianism without military involvement,” “what is the theoretical literature on institutional cooptation,” “which scholars have argued against the transitology framework.” The QA scoring - weighted toward inline OCR at 5, body text at 4, sections at 3, attachment text at 2 - surfaces the entries most likely to contain answers, and the warm-up mechanism that runs background OCR on the top 12 candidates ensures that entries whose content exists only in images are included in the scoring.

The Related Entries panel has surfaced connections across his research in ways that the fragmentary two-product system never could. Notes on cases from different regions, written years apart, that share a structural feature in their political dynamics appear as related when he is working in one of them. “The panel has given me three paper ideas in two years,” he says. “Connections between cases I had studied separately without recognizing their relationship. I consider that a significant research outcome.”

For communications with scholars and civil society contacts in politically sensitive environments, James uses per-entry encryption for the notes he takes about those conversations. The encrypted entries hold only ciphertext on his device. There is no server that holds any version of these notes. Legal process directed at any cloud provider would find nothing, because nothing is there to find.

“What I had before was a two-product system that required constant mental accounting about where things were. What I have now is a single vault where everything I know for my research lives in one place, is searchable at any level of granularity, and is genuinely private in ways that matter for what I actually write about.”

What Every Migration Story Has in Common

Five different people. Five different starting tools. Five different triggers. But reading across these migration stories, a set of common threads emerges that says something about why the moment of switching arrives when it does.

The trigger is almost never a single event. Marcus’s concern about Evernote’s terms of service coexisted with a declining search experience. Priya’s concern about Notion’s AI terms landed on top of years of growing awareness that professional content did not belong in a cloud workspace. David’s source protection concern crystallized in a conversation with a press freedom lawyer. Elena’s HIPAA analysis arrived after years of self-reassurance that iCloud was probably fine. James’s Google assessment came after a decade of using a two-product system that required constant workarounds. In every case, the trigger event arrived after the underlying discomfort had been accumulating for a long time.

The switching cost is consistently smaller than anticipated. Every person in these stories had delayed leaving their previous tool partly because of the perceived cost of migration. The actual migration work ranged from a single afternoon (Elena) to four weeks of intermittent effort (Marcus). In every case, the person described the migration as easier than they had expected and expressed some version of wishing they had done it sooner.

The features they discover in VaultBook exceed what they left. None of these people migrated to VaultBook as a compromise - accepting a feature-poorer tool for better privacy. They migrated and found capabilities they did not have before: vote-based search reranking, inline OCR that makes screenshots searchable, deep PDF indexing that reaches inside attached documents, the Related Entries panel that surfaces connections across large vaults, AI Suggestions that learn from use without sending any data to a server. In several cases, these new capabilities have changed the way the person does their core work.

The privacy architecture changes the quality of the work itself. This thread runs through every story in a way that deserves emphasis. It is not just that VaultBook stores notes more privately. It is that knowing the notes are private changes how the notes are written. Marcus writes more candidly about case strategies because he is not imagining a cloud server indexing the content. David documents source relationships more completely because the documentation is genuinely protected. Elena maintains the clinical formulation notes that are most sensitive without second-guessing whether they should be written down at all. James records his correspondence with politically vulnerable contacts without the ambient anxiety that he was creating an accessible record on a commercial platform. The privacy architecture does not just protect existing notes - it enables notes that would not otherwise have been written.

The vault is genuinely portable. Every person who migrated described the same discovery: the VaultBook vault is a folder. It can be copied to a backup drive with a standard file copy. It can be moved to a new device by moving the folder. It can be opened from an external drive on any computer that has VaultBook.html in the same directory as the libs/ folder. The repository.json file, the details-{id}.md sidecar files, the attachments/ directory, the index/ directory, the versions/ directory - all are open, readable formats that exist as files on the user’s device. There is no vendor lock-in because the data is never on the vendor’s servers to begin with. For people who have experienced the anxiety of a cloud application announcing it is shutting down with thirty days’ notice, or a price change that makes the platform unaffordable, or a terms of service shift that makes continued use uncomfortable, this portability is not a minor feature. It is the essential guarantee that the vault will still exist and be accessible regardless of what any company decides.

The migration is almost always faster than expected. The Import from Obsidian tool accepts standard Markdown files from any source - Evernote ENEX exports converted to Markdown, Standard Notes exports, Obsidian vaults, Bear exports, any text-format output from any application. The tool processes each .md file and creates a corresponding VaultBook entry, preserving the content and heading structure. For the majority of notes in the majority of note-taking applications, this pipeline is sufficient to migrate the full content library in a single session. The cases that require manual effort - notes with complex embedded formatting, heavily structured database-style content, or platform-specific features without a clear equivalent - are a small fraction of a typical library, and they reveal themselves quickly during the migration process rather than requiring discovery after the fact.

The tools we use to think shape what we allow ourselves to think. A note-taking application that lives on a vendor’s server, processed by the vendor’s infrastructure, subject to the vendor’s terms and the legal process directed at the vendor, creates a subtle but real constraint on what the user writes in it. A vault that lives on the user’s device, makes zero background network requests, and protects sensitive content with local AES-256-GCM encryption removes that constraint. The full range of thought - the complete, candid, professionally sensitive, personally vulnerable range that a knowledge-management tool should serve - becomes available.

That is what each person in these stories found when they arrived. Not just better privacy. The ability to use a note-taking tool as a note-taking tool was meant to be used: as a complete, honest record of everything worth knowing.

VaultBook - your personal digital vault. Private, searchable, and entirely under your control.