# Dexi > Dexi (dexi.net) is a lifelong learning system: collect, connect and share notes, bookmarks and feeds in one searchable workspace. Notes organize themselves as you write — #hashtags become tags, @mentions become people and sources, and [[wiki links]] connect ideas — and everything you save is full-text and semantically searchable and part of a private knowledge graph. The app lives at https://app.dexi.net (free to sign up). AI assistants can work with a user's notes directly through the MCP server at https://mcp.dexi.net/mcp (OAuth; instructions at https://dexi.net/mcp). Developer documentation — MCP tools, OAuth details, and the capture API — is at https://docs.dexi.net. # Knowledge Base ## Notes that organize themselves URL: https://dexi.net/kb/notes-that-organize-themselves Topic: Notes Write in a rich-text editor and let structure emerge as you type — #hashtags become tags, @mentions become people and sources, and [[wiki links]] connect ideas. Most note systems make you choose between writing and filing. You either stop mid-thought to pick a folder and add metadata, or you dump everything into one long document you’ll never untangle. Dexi removes the choice: you just write, and the organization happens inline. Type #strategy in a sentence and the note is tagged with it. Mention @jane-goodall and she becomes a trackable source across your whole workspace. Wrap a phrase in [[double brackets]] and it links to the note by that name — creating it the moment you first follow the link. ### Why it matters The value is compounding context. Because tags and mentions are part of the prose itself, every note you write enriches the ones you’ve already written. Six months of reading notes don’t become an archive — they become a network you can pull on from any thread. There’s also no upkeep. Nothing to re-file, no taxonomy to maintain, no “inbox” of unsorted notes accumulating guilt. The editor supports headings, lists, quotes and links, so the same note can be a quick capture or a polished piece of writing. ### Who gets the most out of it - Researchers and graduate students — Literature notes tagged by topic and author cross-reference themselves — the review chapter practically assembles from the network. - Product managers — Meeting notes that mention @customers and #feature-areas become a searchable record of who asked for what, and when. - Journalists and analysts — Sources, claims and themes stay linked from first interview note to published piece. - Consultants — Insights tagged per #client and #industry are reusable on the next engagement instead of buried in a project folder. ### Example: A product manager after a customer call You’ve just finished a call with a customer who had strong opinions about onboarding. While it’s fresh, you open a new note and write: 1. “Call with @acme-corp — their new admins get lost in workspace setup. Strong signal for #onboarding and possibly [[Guided setup wizard]]. They compared us unfavorably to a competitor’s checklist flow.” 2. That’s it. The note is now tagged #onboarding, linked to Acme’s mention page, and connected to your “Guided setup wizard” idea note. 3. Three weeks later, planning the quarter, you open the #onboarding tag and see this call alongside four other customer conversations, two bookmarked articles and your own proposal draft. 4. Clicking @acme-corp shows every note that mentions them — support escalations, the original sales notes, this call — a full account history you never deliberately compiled. You wrote one paragraph. Dexi filed it four different ways. --- ## Search that understands what you meant URL: https://dexi.net/kb/search-that-understands-meaning Topic: Search Full-text search with autocomplete across everything you’ve saved — plus semantic search that finds notes by meaning, even when the words don’t match. Everything you put into Dexi — notes, bookmarks, emailed articles, feed entries — lands in one searchable index. Titles, body text and tags are all covered, with autocomplete that suggests matches as you type. On top of that sits semantic search. Every note gets a meaning-based fingerprint, so you can search for the idea rather than the exact phrase. “That article about founder burnout” will surface a piece titled “Why startup CEOs quit” — no shared keywords required. Every note also has a “similar notes” view that surfaces its nearest neighbors by meaning. When you need precision instead, the search box speaks operators: “quoted phrases” for exact wording, OR for alternatives, -word to exclude, and a query wrapped in slashes — /colou?r/ — runs as a full regular-expression match over titles and text. The ? beside the search box lists them all, with clickable examples. ### Why it matters A capture system is only as good as its recall. The real failure mode of note-taking isn’t losing notes — it’s writing them and never finding them again, because eight months later you don’t remember the words you used. Semantic search removes that dependency: you search with today’s vocabulary and still find last year’s thinking. Because search spans every content type, you stop caring where something came from. Whether an idea arrived as a note you typed, a page you bookmarked, a newsletter you forwarded or an RSS article that flowed in — it’s one query away. ### Who gets the most out of it - Anyone with more than a hundred notes — Keyword search degrades as an archive grows; meaning-based search gets more valuable with every note you add. - Lawyers and compliance teams — Find every note touching a concept — not just the ones that use this year’s terminology for it. - Content creators and writers — “Have I written about this before?” gets a real answer, plus the similar-notes view hands you related material for the piece you’re drafting. - Engineers and technical writers — Search architecture decisions and incident notes by concept when you can’t remember which system they were filed under — or grep them with /regex/ when you’re hunting an exact error code or config key. ### Example: A writer hunting a half-remembered idea You’re drafting an essay and recall reading something months ago about how cities get more productive as they grow while companies get less so. You have no idea what it was called or where you saved it. 1. You type “why cities scale better than companies” into search. 2. Semantic search returns a bookmarked article titled “Geoffrey West on superlinear scaling” — saved five months ago from a newsletter, with none of your query words in the title. 3. From that note, you open Similar notes and find two more: your own reading note on “Scale” and a feed article about organizational overhead. 4. Three sources for your essay, recovered in under a minute, from a search phrased entirely in your own words. --- ## See your knowledge as a graph URL: https://dexi.net/kb/knowledge-graph Topic: Graph Every tag, mention and wiki link builds a private knowledge graph you can explore in 2D or 3D — and traverse to surface connections you didn’t know you’d made. As you tag notes and link ideas, Dexi quietly builds a map: notes, tags, mentions and bookmarks as nodes, every co-occurrence and link as an edge. The graph view renders that map — in 2D for orientation, in 3D when you want to fly through it. The related-concepts traversal takes it further: pick any concept and Dexi walks the graph outward, showing which ideas cluster together and which single note bridges two otherwise separate territories of your thinking. ### Why it matters Lists show you what you were looking for; graphs show you what you weren’t. Clusters reveal your true centers of gravity — which themes actually dominate your reading and writing, as opposed to the ones you think do. Sparse regions show where a research area is thin. Bridge nodes often turn out to be the interesting ideas: the note that connects two clusters is frequently the seed of original work. It’s also the fastest honest overview of a body of work. A glance at the graph tells you more about six months of research than scrolling a chronological list ever will. ### Who gets the most out of it - Academics and PhD students — Spot the gap in the literature review — visible as the missing edge between two clusters you’ve been reading around. - Strategy and innovation teams — Cross-domain connections are the raw material of strategy; the graph makes them visible instead of accidental. - Nonfiction authors — Watch a book’s structure emerge from research clusters, and find the through-line that connects the chapters. - Investors and market researchers — See which companies, technologies and theses are converging — the bridge notes often point at where a market is heading. ### Example: An investor reviewing a year of research You’ve spent a year saving notes and articles about developer tools and, separately, about healthcare IT. They’ve always felt like unrelated interests. 1. You open the graph view. Two dense clusters appear — #devtools and #healthtech — exactly as expected. 2. But one node sits between them: a note on “FHIR APIs”, connected to both clusters through a bookmarked article and two of your own notes about developer experience in regulated industries. 3. You run the related-concepts traversal from that node and watch it pull in #compliance, @a16z’s piece you saved last spring, and a feed article on health-data startups. 4. You didn’t plan a thesis about developer platforms for healthcare data. The graph shows you’ve been building one for months. --- ## Capture any page, from anywhere URL: https://dexi.net/kb/capture-the-web Topic: Bookmarks Save pages in one click with the web clipper extension (Chrome, Brave, Arc, Orion, Vivaldi) or bookmarklet, from your phone with the iOS share sheet, or import your existing Chrome bookmarks wholesale. Dexi meets you where you already read. The web clipper extension (Chrome, Brave, Arc, Orion and Vivaldi — or the bookmarklet, in any browser) saves the page you’re on — title, link, favicon and any text you’ve selected — with a keyboard shortcut or a right-click. On iPhone, a share-sheet shortcut posts pages straight from Safari or any app. Already have years of bookmarks? Import your Chrome bookmarks in one step and they arrive with titles and favicons intact. Coming from Pocket? Upload the CSV or ZIP you exported before it shut down and your Pocket tags become Dexi tags. Every saved page becomes a note, which means it’s taggable, searchable — including semantically — and part of your knowledge graph like everything else. ### Why it matters Browser bookmarks are where links go to die: unsearchable beyond their titles, unorganized past the first month, trapped in one browser. Saving into Dexi instead means the page joins your actual thinking — you can tag it #competitive-analysis on the way in, find it later by what it was about, and see it connected to your notes on the same topic. The capture cost is low enough that you’ll actually do it. One click, an optional note, done — no app-switching, no copy-pasting URLs into a document you’ll lose. ### Who gets the most out of it - Market and competitive researchers — Pricing pages, launch posts and job listings saved in the moment become a longitudinal record of a market. - Designers and creative professionals — A tagged reference library of inspiration that’s searchable by concept, not buried in browser folders. - Recruiters and talent teams — Profiles and portfolios captured from the phone between meetings, tagged by role and pipeline. - Anyone switching from browser bookmarks — The Chrome import turns a decade of dead links into a searchable, taggable library in one step. ### Example: A competitive researcher’s Tuesday morning A competitor announced pricing changes overnight, and links are flying around Slack. 1. On the competitor’s pricing page, you select the paragraph describing the new enterprise tier and hit the extension shortcut. The popup opens with the page and your selected text pre-filled; you add #competitor-acme #pricing and save without leaving the tab. 2. Their CEO’s thread explaining the change: right-click, save to Dexi, same tags. 3. On the train, an analyst’s take lands in your feed reader — you share it to Dexi from your phone with the iOS shortcut. 4. That afternoon you open the #competitor-acme tag: today’s three saves sit alongside eight months of launch notes and your own analysis, ready to become a briefing. --- ## Forward an email, get a note URL: https://dexi.net/kb/email-to-note Topic: Email-in Send or forward anything to your personal save address and it becomes a note — newsletters, receipts, ideas from your phone — filed and instantly searchable. Every Dexi account has a personal save address — save-xxxxx@my.dexi.net, shown in Settings → Save by email. Send or forward any message to it and it lands in your workspace as a note, with the subject as its title and the body as its content. Hashtags in the message become tags, just as they would in the editor. The address itself is the key, so mail from any sender counts: forward by hand, set up a forwarding rule for the newsletters you always keep, or Bcc it on mail you send. Nothing to install — if your inbox can forward, it can file. If the address ever leaks, regenerate it in Settings and the old one stops working. ### Why it matters Email is where an enormous amount of valuable reading actually happens — newsletters, industry digests, long threads with collaborators — and where it all disappears. Forwarding the good ones to Dexi separates “worth keeping” from the inbox churn, and makes them full citizens of your knowledge base: searchable by meaning, taggable, connected to your notes. It’s also the lowest-friction capture channel there is for ideas. An email to yourself becomes a note in your system rather than one more unread message in your inbox. ### Who gets the most out of it - Newsletter power-readers — The five best issues a week get forwarded and become searchable research; the rest stay in the inbox where they belong. - Executives and busy operators — Capture from the inbox — where you already live — without adopting a new tool or workflow. - Real estate and field professionals — Listings, inspection reports and client emails filed by forwarding, straight from the phone between appointments. - Teams doing customer research — Notable customer emails forwarded with a #voice-of-customer tag build an evidence base over time. ### Example: A founder’s newsletter habit You read a dozen newsletters a week. Most are skimmed and forgotten — but this morning’s issue has a sharp breakdown of usage-based pricing you know you’ll want when your own pricing review starts next quarter. 1. You hit forward, type your save address (save-k7x2m@my.dexi.net — it autocompletes after the first time), and add one line above the quoted issue: “Best breakdown of usage-based pricing I’ve seen #pricing #saas”. 2. Seconds later it’s a note in Dexi — your comment on top, the full issue below, tagged #pricing and #saas. 3. Next quarter, preparing the pricing review, you search “usage-based pricing”. The newsletter is there — alongside two bookmarked articles and your own notes from last year’s attempt. 4. The newsletter itself was deleted from your inbox months ago. The knowledge wasn’t. --- ## Bring your sources to you URL: https://dexi.net/kb/rss-subscriptions Topic: Feeds Subscribe to RSS feeds and let new articles flow into your workspace hourly — searchable, taggable and connected to everything else you know. Subscribe to any RSS or Atom feed — blogs, news sites, journals, podcasts’ show notes — and Dexi checks for new entries every hour. Each article arrives as content in your workspace: readable, taggable and included in both full-text and semantic search. You can also share subscriptions with other Dexi users, so a team or reading group can run from the same set of sources. ### Why it matters A feed reader tells you what’s new; Dexi tells you what’s new and how it relates to what you already know. Because incoming articles live in the same index as your notes and bookmarks, a semantic search for a topic surfaces this morning’s article next to your own six-month-old analysis of the same idea. Reading becomes accumulation instead of consumption. It also ends the tab-hoarding workflow. Sources you trust flow in on their own schedule — you stop patrolling twelve sites and let the system do the rounds. ### Who gets the most out of it - Analysts and researchers on a beat — Coverage of a sector becomes automatic; analysis stays connected to the primary sources it draws on. - Engineers tracking their ecosystem — Release blogs and engineering feeds land in the same searchable space as your own architecture notes. - Academics following journals — New papers from journal RSS feeds arrive pre-connected — via search and the graph — to your existing literature notes. - Editorial and content teams — A shared subscription set gives the whole team one pool of raw material for planning and curation. ### Example: An analyst covering the energy sector You cover grid-scale storage. You subscribe Dexi to a handful of sources: two industry blogs, a regulator’s news feed, and a research group’s publications feed. 1. Articles arrive on their own, every hour, while you do other work. 2. Wednesday, a regulatory filing about interconnection queues appears. Reading it in Dexi, you tag it #interconnection and add a one-line note about why it matters. 3. A search for “interconnection delays” now returns the filing, a blog post from last month, and your own client memo from Q1 — one thread of evidence across three different sources. 4. When a client asks “what’s changed this quarter?”, the answer is a tag away instead of a frantic browser-history reconstruction. --- ## Digests on your schedule URL: https://dexi.net/kb/scheduled-digests Topic: Digests Get a recurring email digest of notes matching the tags you choose — a built-in review habit for you, or a curated briefing for anyone else. A digest is a standing order: “every Friday at 8am, email me everything new under #reading-list”. You pick the tags, the schedule and the recipient, and Dexi assembles and sends the email — in your timezone, without you thinking about it. Digests work for any tag and any cadence, and you can run several at once: a Monday digest of #team-updates, a weekly #inspiration roundup, a monthly review of #ideas. ### Why it matters Capture without review is hoarding. The digest closes the loop: material you saved during the week comes back to you at a moment you chose for reading rather than collecting. It’s spaced repetition for your own research, run by the system instead of your willpower. Pointed at someone else, a digest becomes effortless publishing. Tag as you save, and a curated briefing goes out on schedule — no newsletter tooling, no “weekly roundup” doc to remember to write. ### Who gets the most out of it - Chiefs of staff and team leads — Tag notable items #exec-briefing during the week; the Friday digest to your CEO writes itself. - Lifelong learners — A weekly review email turns saved-and-forgotten into actually-revisited. - Educators and program leads — A #course-readings digest delivers curated material to students on the course’s rhythm. - Community and DevRel managers — Ecosystem news tagged through the week becomes a polished community roundup on schedule. ### Example: A chief of staff’s Friday briefing Your CEO wants a weekly pulse on the market but has no time to gather it. You set up a digest: tag #ceo-brief, every Friday, 7am, to their inbox. 1. Through the week you work normally — reading feeds, saving bookmarks, taking meeting notes. Anything briefing-worthy gets the extra tag #ceo-brief as you save it. Total added effort: seconds per item. 2. A competitor’s funding announcement from Tuesday, a regulatory note from your RSS feeds, and your summary of a key customer call all pick up the tag. 3. Friday 7am, the digest lands in the CEO’s inbox: this week’s tagged items, titles and summaries, formatted and linked. 4. The briefing that used to eat your Thursday evening now assembles itself from work you were already doing. --- ## Publish your research as a feed URL: https://dexi.net/kb/published-feeds Topic: Publishing Turn any slice of your workspace into a public page with its own RSS feed — curated by tag, updated as you save, subscribable by anyone. A published feed is the outbound half of Dexi. Choose the tags that define it — say #ai-safety — and Dexi gives you a public page and a standard RSS feed of matching notes. Save and tag as you normally would; the feed updates itself. Readers don’t need a Dexi account: the page is a link you can share anywhere, and the RSS feed works in any reader or automation that speaks RSS. Everything else in your workspace stays private — only what you deliberately tag into the feed is published. ### Why it matters Curation is valuable and chronically underpublished, because the tooling makes it a second job — a newsletter platform, a blog, a posting habit. A published feed reduces publishing to tagging: if you’re already collecting the best material on a topic, sharing it costs nothing more. A consistently good curated feed is also a reputation engine. It demonstrates judgment — what you chose and what you skipped — which is precisely the skill clients, employers and audiences are trying to evaluate. ### Who gets the most out of it - Independent consultants and fractional executives — A curated feed on your specialty is ongoing proof of expertise that markets you between engagements. - Developer advocates — The ecosystem-news-worth-knowing feed your community wants, produced as a byproduct of your own reading. - Research leads and librarians — Publish a monitored slice of the literature for your department without running a mailing list. - Investors — A public reading feed on your thesis area attracts exactly the founders you want to meet. ### Example: A consultant building visibility You’re an independent consultant in supply-chain resilience. You already save everything good on the topic — so you publish it. 1. You create a published feed from the tag #supply-chain-resilience and put the public link in your email signature and LinkedIn profile. 2. Your normal reading now doubles as publishing: a bookmarked port-congestion analysis, a forwarded newsletter piece, a short note of your own — each tagged on the way in, each appearing on your public page. 3. A prospective client subscribes to the RSS feed in their reader. For two months, your judgment shows up in their morning reading. 4. When their resilience audit finally gets budget, the shortlist is one name long. --- ## Let Claude work your notes URL: https://dexi.net/kb/use-with-claude Topic: AI Dexi’s built-in MCP server connects your workspace to Claude, ChatGPT, Goose and other AI assistants — search, read, create and update notes by conversation. Dexi ships an MCP (Model Context Protocol) server, the open standard AI assistants use to work with your tools. Connect it once — from claude.ai, Claude Desktop, Claude Code, ChatGPT (developer mode), Goose or any MCP client — and your assistant can search your notes by keyword or meaning, read them, list your tags, find similar notes, and create, update or delete notes on your instruction. The connection is scoped and revocable: you approve exactly what an assistant can do — and what it can touch: everything, a single folder, a single tag, or a tag within a folder. A restricted assistant sees only notes inside that boundary, and notes it creates are filed there automatically. You can change the boundary or disconnect any app instantly in Settings → Connected apps. Your password is never shared. ### Why it matters This turns your workspace from a place you search into a research partner you can brief. “What have I saved about pricing experiments?” becomes a question you ask in conversation — and the answer draws on your notes, not the open internet. Summaries, syntheses and drafts are grounded in what you actually collected. It works in both directions: the assistant can file things too. A conclusion reached in a Claude conversation can be saved as a note — tagged, searchable, part of the graph — without you switching apps. ### Who gets the most out of it - Writers and researchers using AI daily — Drafts and syntheses grounded in your own source notes instead of the model’s general knowledge. - Developers living in Claude Code — Search architecture decisions and log new ones from the terminal, mid-task, without breaking flow. - Knowledge workers with a serious archive — A thousand notes become genuinely conversational — ask, don’t excavate. - Automation builders — Any MCP-capable agent can read from and write to your knowledge base as a building block in larger workflows. ### Example: A researcher synthesizing a year of notes You’ve spent a year collecting material on remote-work productivity — notes, forwarded studies, bookmarked articles. A client wants a briefing by Thursday. 1. In claude.ai, you’ve connected Dexi as a custom connector (Settings → Connectors → https://mcp.dexi.net/mcp) and approved read/write access. 2. You ask: “Search my Dexi notes for everything about remote work and productivity, and summarize the main findings with disagreements called out.” 3. Claude runs semantic searches, reads the relevant notes, and returns a synthesis — noting that your saved studies genuinely disagree on the effect of hybrid schedules, with the specific notes cited. 4. You refine it in conversation, then say: “Save this as a new note titled Remote work briefing — client draft, tagged #remote-work #client-x.” It’s in your workspace, in the graph, done. --- ## Your knowledge, yours to take URL: https://dexi.net/kb/own-your-data Topic: Export Export everything — notes as Markdown, bookmarks as CSV, the whole workspace as a ZIP — anytime, in formats every other tool can read. Everything you put into Dexi comes back out. Export your notes as Markdown files, your bookmarks as CSV, or the whole workspace as a ZIP archive — self-serve, whenever you want, no support ticket required. The formats are deliberately boring: Markdown and CSV are readable by humans, importable by virtually every tool, and will still open in fifty years. ### Why it matters A knowledge base is a decades-long bet, and lock-in is how that bet goes bad. Real export — complete, standard formats, on demand — means the switching cost stays near zero forever, and you can commit to the system precisely because you’re free to leave it. Exports also serve the practical present: keep an independent backup on your own storage, feed your notes into external analysis or AI pipelines, or hand a Markdown bundle of project research to a collaborator who’s never heard of Dexi. ### Who gets the most out of it - Anyone burned by a dead platform — If you’ve lost an archive to a shutdown or an acquisition before, exportability is the first feature you check. - Regulated industries — Legal, finance and healthcare teams can meet retention and audit obligations with independent archives. - Developers and tinkerers — Markdown and CSV pipe cleanly into scripts, static sites and analysis tools. - Consultants and agencies — Deliver the research trail to a client as portable files at the end of an engagement. ### Example: A consultant closing out an engagement A six-month engagement is wrapping up, and the client’s counsel wants the supporting research preserved on their side. 1. You open Export in Dexi and download the ZIP archive of your workspace. 2. The client-project notes — Markdown, with tags and titles intact — go into the final deliverable package; the bookmarks CSV documents every source consulted. 3. The client’s team imports the Markdown into their own wiki without conversion. No licenses, no viewers, no proprietary formats. 4. You keep the full archive in your own backups — the engagement’s knowledge outlives the engagement, on both sides. --- ## Tags that stay useful URL: https://dexi.net/kb/tags-that-stay-useful Topic: Tags Tags in Dexi come from your writing, show up with live counts on the tags page, and autocomplete as you type — so your taxonomy stays real instead of rotting. Tags fail in most tools for a predictable reason: they’re created in a hurry, never seen again, and multiply into near-duplicates nobody remembers. Dexi treats tags as a living part of the workspace instead of write-only metadata. Every #hashtag you type becomes a tag, and the tags page shows them all with counts — a live census of what you actually think about. When you start typing a tag, autocomplete suggests existing ones, nudging you toward #onboarding instead of accidentally minting #on-boarding. And because bookmarks, emailed notes and feed articles are taggable too, one tag collects a topic across every format. ### Why it matters The counts are the quiet superpower. A tag list with numbers is a map of your attention: the topics you thought were side interests but have forty notes, and the “priorities” with three. That’s information you can act on — and it’s generated for free by writing normally. Autocomplete keeps the vocabulary converging rather than diverging. The longer you use Dexi, the more coherent your tag system gets — the opposite of what happens in most tools. ### Who gets the most out of it - Knowledge-management enthusiasts — A self-maintaining taxonomy delivers what elaborate folder hierarchies promise and never keep. - Content strategists and editors — Tag counts show which themes are over- and under-served before planning the next quarter. - Product and UX researchers — Recurring themes across interview notes become visible — and countable — evidence. - Anyone returning after a break — The tags page is the fastest re-orientation there is: here’s what you were thinking about, ranked. ### Example: A UX researcher’s quarterly review You’ve run twenty user interviews this quarter, each written up as a note with tags applied while typing — #navigation, #pricing-confusion, #mobile, whatever came up. 1. At quarter’s end you open the tags page. #pricing-confusion has 11 notes — you’d have guessed four. #mobile has 3, despite mobile being the roadmap’s headline. 2. You click #pricing-confusion and get all eleven interview notes together, plus two bookmarked competitor teardowns that picked up the same tag along the way. 3. While writing the summary, you type #pri… and autocomplete offers the exact tag — no #pricing vs #pricing-confusion fork to reconcile later. 4. Your quarterly readout leads with the finding the tag counts surfaced: users aren’t confused by the product, they’re confused by the pricing. --- ## Folders, when you want them URL: https://dexi.net/kb/folders-when-you-want-them Topic: Folders A flat set of personal folders for the handful of things you’re actively working on — file any note, bookmark or emailed article in one click, and browse a folder like it’s its own workspace. Dexi’s organization has always been emergent: tags come out of your writing, links come out of your thinking, and nothing needs filing. That works beautifully for a body of knowledge — and less well for the four or five things you happen to be working on right now. Folders are for those. They’re deliberately plain. Create a folder in the sidebar, put a note in it, click the folder to see what’s inside. There’s no nesting to design and no hierarchy to maintain — just a flat list with live counts, plus an Unfiled row for everything that hasn’t been filed. A note lives in at most one folder, which is what makes “where is it?” a question with one answer. ### Why it matters The point isn’t to replace tags — it’s that folders and tags answer different questions. A tag says what a note is about, and one note can carry five of them; a folder says what it belongs to, and picking one is a decision. Client work, a thesis chapter, a launch, a house move: these are containers, not topics, and forcing them into a tag vocabulary has always been the awkward part of tag-only systems. Now #pricing can appear in three projects while each project keeps its own drawer. Folders cover everything, because everything in Dexi is a note — a bookmarked page, a forwarded email and an RSS article can all be filed alongside your own writing. Captured items arrive unfiled by default, so Unfiled works as an inbox you can triage when you feel like it, or ignore entirely. Filing is metadata-only: moving a note doesn’t touch its content or bump its “updated” date, so tidying up never disturbs your recent-work view. Deleting a folder keeps every note in it — they simply become unfiled again. And your filing is yours alone: even inside a shared team, teammates see the notes you share but never the folders you keep them in. ### Who gets the most out of it - Freelancers and consultants — One folder per client keeps engagements cleanly separated, while tags still let insights travel across all of them. - Students and researchers — A folder per course, chapter or paper gives each piece of work a boundary — with the whole library still searchable as one thing. - People with a capture habit — Bookmarks, forwarded newsletters and feed articles pile up in Unfiled, ready to sort into projects in one pass instead of at capture time. - Anyone who never got on with tags — Folders are the familiar mental model — start here, and let tags accumulate underneath as a bonus rather than a prerequisite. ### Example: A freelance designer running three clients at once You’re mid-flight on three engagements, and your workspace is one long stream of call notes, saved articles and forwarded briefs. You spend two minutes setting up folders. 1. In the sidebar you create Acme Rebrand, Northwind Site and Personal — three folders, no hierarchy. Underneath them, Unfiled shows the 90-odd items that have accumulated. 2. You open Unfiled and work down the list. Each note’s menu has “Move to folder”, so a client brief, two competitor teardowns you bookmarked and a forwarded email from Acme’s marketing lead all land in Acme Rebrand without leaving the list. 3. Before a call, you click Acme Rebrand in the sidebar and get the whole engagement — your notes, their emails, the articles you saved — as its own view. The count next to the folder name tells you it’s grown to 34 items since kickoff. 4. Filters still stack on top: inside Acme Rebrand you filter to #typography, then run a semantic search for “what did they say about the wordmark” — scoped to that client, not your entire workspace. 5. Meanwhile #typography stays global. Clicking it from the tags page shows every typography note across all three clients and your personal reading — the cross-client pattern you’d lose in a folders-only tool. 6. When the Northwind project wraps, you delete the folder. Every note survives and quietly returns to Unfiled, still tagged, still searchable, still in your graph — a closed project rather than a deleted one. Folders for what you’re working on, tags for what you know. Two minutes of setup, and neither one has to fight the other. --- ## Properties and bases: a table view of your notes URL: https://dexi.net/kb/properties-and-bases Topic: Bases Write key:: value lines in any note and Dexi turns them into typed properties. A base is a saved table, card, list or map view over the notes that match a filter, with formulas, grouping and column summaries. Obsidian-compatible: .base files import directly. Sometimes a stack of notes wants to be a table. A reading list with a status and a rating. Job applications with a company, a stage and a next date. Recipes with a cuisine and a cook time. Dexi has never had a properties panel or a schema to set up first, and it still doesn’t: a property is a line you type. Write status:: reading on its own line, anywhere in the body, and the note now has a property called status with the value reading. due:: 2026-09-10 becomes a date, rating:: 4 a number, authors:: [Kahneman, Tversky] a list, ref:: [[Thinking, Fast and Slow]] a link. Dexi reads the type from the shape of what you wrote, so there is nothing to declare. A base is where those lines pay off: a saved view — table, cards, list or map — of every note that matches a filter, sorted or grouped how you like, with the properties (and any formulas you define) as columns. ### Why it matters Properties are prose, which keeps them cheap. There’s no form to fill in and no field to add to a template before you can use it, and the line stays visible in the note where you wrote it, so a note read on its own still makes sense. Keys are forgiving — Due Date:: and due_date:: are the same property — and a value that looks like a date, a number, true/false, a #tag or a [[wiki link]] is treated as one. Property lines stay in full-text and semantic search, so author:: Kahneman is still found by searching for Kahneman. Lines that only look like properties — std::vector, a URL, a quoted > reply:: — are left alone. A base is a saved question over your notes. Its filters can use your own properties and the things Dexi already knows about every note: title, tags, links, folder, source (typed, bookmark, email, feed), domain, created and updated dates, whether a teammate shared it. Conditions combine with and, or and not, sorts stack up to five deep, and one base can hold several views — the same reading list as a table sorted by rating and as cards showing each page’s preview image. Bases run over everything you can see, including notes teammates have shared with you, and they’re personal: nobody else sees your bases or how you’ve set them up. Every base is also available to your AI assistant — a connected Claude can run “my reading list” or “what’s due this week” through the MCP server instead of reconstructing the filter. If you’ve built Bases in Obsidian, the filter language is the same subset you already know and a .base file imports as-is, formulas included (a formula like rating * 2 or due < today() becomes a column, a sort key or a filter, exactly as in Obsidian); views group by any property and summarize columns (average, sum, earliest, count…) exactly as in Obsidian; there are list and map layouts beside table and cards (a coordinates:: 37.77, -122.41 property puts a note on the map); only custom summary formulas aren’t here yet, and an import tells you exactly what it left out rather than dropping it silently. ### Who gets the most out of it - Readers with a queue — A reading list is the canonical base: a status, a rating and a date finished, next to the page’s own description and preview image that came in with the bookmark. - Job seekers and freelancers — Company, stage, next step, contact — one line each in the note you already keep, and a table of open applications or proposals that sorts itself by next date. - Students and researchers — Papers with an author, year, venue and verdict become a literature table you can filter to “unread”, sort by year, and export alongside your Markdown. - Obsidian users — The same key:: value syntax and the same Bases filter language, with your .base files importing directly — running over bookmarks, forwarded newsletters and feed articles as well as typed notes. ### Example: A reading list that keeps itself You bookmark a lot of long articles and lose track of which ones you’ve actually read. You decide to give the queue a table. 1. You open a saved article and add two lines at the top of the note: status:: unread and topic:: #pricing. The editor highlights the status:: prefix, so you can see it was recognized as a property. From now on each article you clip gets the same two lines. 2. You already have the list filtered to #book bookmarks, so you click Save as base above it and name the new base Reading list — it opens as a table with that filter in place (or start from the plus next to Bases in the sidebar). Under Filter you add one condition — status is not done — and under Columns you pick Title, status, topic, Domain and Updated. 3. You add a second view, choose Cards, set the image property to Preview and pick a card size. The same queue is now a wall of article thumbnails. 4. Back in the table, Group & sort groups the rows by topic, and from the rating column’s menu you pick Summarize → Average: each topic gets a header with its count and average rating, and a summary row totals the queue. Under Formulas you add overdue = due < today(), and it shows up as a column you can sort and filter like any property. 5. Back in the table you sort by Updated, newest first, then by topic, and set a limit of 50 so the view stays a queue rather than an archive. Each change saves to the base as you make it. 6. When you finish an article you change its line to status:: done and it drops out of the view. Add rating:: 4 beneath it and rating appears in the column picker for every view, ready to sort on. 7. In Claude, connected through the MCP server, you ask what’s on your reading list about pricing. It runs the Reading list base and answers from the entries tagged #pricing, properties included. 8. Months later you export the base as a .base file — the format Obsidian uses — to keep next to your Markdown export, or to open the same view in a vault. Two typed lines per note and one saved view, and the queue takes care of itself. --- ## Connecting notes: wiki links and backlinks URL: https://dexi.net/kb/wiki-links-and-backlinks Topic: Links Wrap a phrase in [[double brackets]] to connect notes. Every link is a backlink on the other side and an edge in your graph — web-based, free for 1,000 notes. Connecting notes is the point of a modern notes app, and in Dexi it happens while you write. Wrap a concept in [[double brackets]] — [[spaced repetition]], [[Q3 pricing review]], [[the maintenance trap]] — and it becomes a wiki link to the note by that name. If that note doesn’t exist yet, following the link creates it, so you can sketch the shape of an idea before filling it in. Links work in both directions. The note you linked to shows the link as a backlink, so a concept note collects every place you have referred to it without you filing anything, and each link is an edge in your knowledge graph — the connections you make while writing are the same ones the graph view and related-concepts traversal explore later. Ordinary hyperlinks work alongside wiki links: the same sentence can point at a web page and at your own note about it. ### Why it matters Linking while writing captures something tags can’t: the specific, directional claim that this idea relates to that one. A tag says “this is about pricing”; a link says “this argument depends on [[willingness to pay]]”. Over time those little claims accumulate into a personal wiki — your own encyclopedia of the concepts you work with, written in your own words. That is what people mean by a note-taking app with backlinks, and the connected notes are what make search and the graph worth having. Because Dexi is a web app, the linked notes are the same in every browser — nothing to install, no sync to configure, and the first 1,000 notes are free. The create-on-first-reference behavior also changes how you plan: you can write an outline whose bracketed sections don’t exist yet, and the empty links become a visible to-do list for your thinking. ### Who gets the most out of it - Zettelkasten and evergreen-notes practitioners — Concept notes linked at the moment of writing are the method, with no plugin required. - Novelists and worldbuilders — Characters, places and plot threads become a story bible that cross-references itself. - Engineers keeping decision logs — Each design decision links to the constraints and prior decisions it rests on — an audit trail of reasoning. - Students building durable understanding — Linking new concepts to known ones is the encoding step that makes material stick. ### Example: An engineer writing a design doc You’re writing up a decision to move billing to event-driven processing. As you draft the note, you link the load-bearing concepts: 1. “We’ll adopt [[event-driven billing]] despite the complexity cost, because [[invoice latency]] is now our top enterprise complaint. This reverses [[2024 billing simplification]].” 2. [[invoice latency]] and [[2024 billing simplification]] already exist — the new note is now wired into their histories, and each of them links back here. 3. [[event-driven billing]] doesn’t exist yet. You follow the link, and the empty note is created — you drop in two bookmarked articles and a paragraph of context while it’s fresh. 4. A year later, a new teammate asks why billing is event-driven. You send one link; the backlinks walk them through the entire decision lineage. --- ## Everything becomes a note URL: https://dexi.net/kb/one-searchable-workspace Topic: Workspace Bookmarks, forwarded emails and feed articles all land as notes — one format, so tags, search, the graph, digests and export work identically on everything you save. Dexi has one load-bearing design decision: every capture channel produces the same thing. A page you bookmark, a newsletter you forward, an RSS article that flows in — each becomes a note, exactly like the ones you type. That single format is why the rest of the system composes. Tags work on everything. Full-text and semantic search cover everything. The knowledge graph connects everything. Digests can include everything, published feeds can share everything, and export takes everything with you. ### Why it matters Most people’s research is fragmented by arrival channel: links in the browser, newsletters in the inbox, articles in a feed reader, thoughts in a notes app — four silos, four searches, zero cross-references. Collapsing them into one format means the question “where did I see that?” loses its power. You saw it in Dexi, whatever door it came through. It also means every feature you learn applies everywhere. There’s no “bookmarks can’t be tagged” or “emails don’t show in search” asterisk — one mental model covers the whole workspace. ### Who gets the most out of it - Multi-channel readers — If your inputs span browser, inbox and feeds, unification is the single biggest quality-of-life gain. - Researchers who cite sources — Typed analysis and captured source material live in one graph, so claims stay attached to evidence. - Teams standardizing on a workflow — One format means one set of conventions to teach, whatever tools people capture from. - Tool-fatigued knowledge workers — Replaces the read-later app, bookmark manager and newsletter archive with one system. ### Example: A speaker assembling a conference talk You’re giving a talk on AI coding agents in six weeks. Material for it has been arriving through every channel for months, all tagged #ai-agents on the way in. 1. You open the #ai-agents tag: nine bookmarked papers and blog posts, four forwarded newsletter issues, a dozen feed articles from engineering blogs, and seven notes of your own commentary — one list, one format. 2. A semantic search for “agents failing at long-horizon tasks” pulls the three most relevant items — one from each original channel, not that you can tell anymore. 3. You draft the outline as a note, wiki-linking the strongest sources directly into each section. 4. The bibliography slide takes ten minutes: every source is already in one place, with its link, title and your margin notes attached. --- ## Merge fragments into one note URL: https://dexi.net/kb/merge-notes Topic: Notes Captured the same idea three times in three places? Merge notes combines them — content and tags together — so quick capture never turns into permanent clutter. Frictionless capture has a byproduct: duplicates. You jot a thought on Monday, bookmark an article about it Wednesday, and email yourself a better version of the same idea the following week. Three notes, one idea. Merge fixes this without ceremony: pick the notes that belong together and combine them into one, keeping their content and their tags. The fragments disappear; the consolidated note stays connected to everything the originals were. ### Why it matters The deeper value is that merge makes low-friction capture safe. If consolidating later is cheap, you never have to interrupt a thought to check “do I already have a note about this?” — you write now and tidy on your own schedule, which is exactly the capture behavior that makes a knowledge base actually get used. It also mirrors how understanding develops. Ideas legitimately start as scattered fragments; merging is the moment you recognize they were one idea all along. The tool supports the synthesis instead of punishing the mess that precedes it. ### Who gets the most out of it - Fast, messy capturers — If you optimize for never losing a thought, merge is the cleanup crew that makes the habit sustainable. - Writers moving from fragments to drafts — Consolidating scattered observations into one note is literally the first step of drafting. - Long-time users with legacy clutter — Years of imports and quick saves inevitably contain duplicates worth folding together. - Meeting-heavy roles — Four notes from four meetings about the same decision become one authoritative record. ### Example: A writer consolidating book notes Over two months you’ve accumulated three separate notes about the same book: highlights typed after reading, a bookmarked review you agreed with, and a late-night emailed thought titled “that maintenance idea again”. 1. Preparing to write your own review, you search the book’s title and all three fragments surface. 2. You merge them into one note. The highlights, the review link and the late-night thought are now one document, carrying the union of their tags — #reading-notes, #maintenance, #essay-ideas. 3. Reading the merged note top to bottom, the through-line you’d been circling for two months is suddenly obvious — the late-night email was the thesis all along. 4. You retitle it, add an outline at the top, and the review drafts itself from material you already had. --- ## Automate capture with API tokens URL: https://dexi.net/kb/save-with-api-tokens Topic: Automation Mint a personal access token and anything that can make an HTTP request can save bookmarks into your workspace — the iOS share sheet, scripts, or your own tools. In Settings → API tokens you can mint a personal access token — a dxi_-prefixed secret shown once and revocable anytime. A token authenticates exactly one thing: saving bookmarks. That narrow scope is deliberate — a token in a phone shortcut or a script can add to your library, but it can’t read, change or delete anything. The flagship use is the iPhone share sheet: an iOS Shortcut holding your token turns “share → Dexi” into a save from Safari, Twitter, your RSS reader — any app that can share a link. But anything that speaks HTTP works: a shell alias, a step in an automation platform, a cron job. ### Why it matters Capture tools live or die on being available at the moment of encounter, and a lot of encounters happen on the phone or inside other tools. Tokens extend one-click saving beyond the web clipper to effectively everywhere. The write-only scope changes the risk calculus of automating. You can put a token in an iOS Shortcut, a coworker-visible script or a third-party automation with a clear worst case: if it leaks, someone could add bookmarks to your account until you revoke it — not read a decade of private notes. ### Who gets the most out of it - Mobile-first readers — Most discovery happens on the phone; the share sheet makes the phone a first-class capture device. - Developers — A curl one-liner or shell alias saves links from the terminal without touching a browser. - Automation builders — Zapier-style workflows and scripts can route links from other services into your library. - Security-conscious users — Hashed at rest, shown once, single-purpose, revocable per token — automation without handing out your password. ### Example: A developer’s phone-to-workspace pipeline You read mostly on your phone — Mastodon, newsletters, a feed reader — and links you mean to keep die in a self-messaged chat thread. 1. In Dexi’s Settings → API tokens you mint a token named “iPhone shortcut” and copy it — it’s shown this once. 2. You build a two-step iOS Shortcut: receive a URL from the share sheet, then POST it to Dexi’s bookmark-save endpoint with the token in the header. Add it to the share sheet and you’re done. 3. From then on, saving from any app is share → Dexi. The page lands in your workspace with its title, ready to tag next time you’re at a desk. 4. Months later the shortcut misbehaves after an iOS update; you revoke just that token in Settings — your password, sessions and other tokens are untouched. --- ## Share your sources, not just your findings URL: https://dexi.net/kb/share-feed-subscriptions Topic: Collaboration Share an RSS subscription with a colleague by email — they connect it to their own workspace, and you’re both building on the same stream of sources. Research teams usually share conclusions — the memo, the deck — while the sources that produced them stay siloed in individual feed readers. Dexi lets you share the subscription itself: pick one of your feeds, enter a colleague’s email, and they get an invitation to connect that source to their own workspace. If they’re on Dexi, one click adds the feed and the articles start flowing into their workspace like any other subscription. If they’re not, the invitation shows them what they were invited to and gets them started. ### Why it matters Shared sources compound differently than shared conclusions. When two people read the same stream, every conversation starts from common ground — “did you see the piece on Tuesday?” works. And because each person tags and annotates in their own workspace, you get independent readings of the same material instead of one person’s filter. It’s also the lightest possible onboarding for a topic. Handing a new teammate your five best feeds transfers months of source-curation judgment in a minute. ### Who gets the most out of it - Research and analyst teams — A shared source base with independent annotation is how you get both alignment and diverse readings. - Managers onboarding new hires — “Here are the feeds this role runs on” beats a stale links doc in the wiki. - Reading groups and study partners — Everyone subscribed to the same journals discusses from the same page. - Mentors and educators — Sharing your sources teaches the judgment behind them — where to look, not just what was found. ### Example: Onboarding a new analyst A new analyst joins your market-intelligence team on Monday. The team’s real edge is its reading list — the feeds you’ve curated over two years. 1. You open your subscriptions and share the five that matter — two industry blogs, a regulator’s feed, a research group, one indispensable newsletter’s RSS — to her email. 2. She accepts the invitations and connects the feeds; by Tuesday, the same articles arriving in your workspace are arriving in hers. 3. She tags and annotates in her own workspace as she reads — by week three, her fresh-eyes reading of the regulator’s feed catches a filing pattern the team had stopped noticing. 4. Standing team habit from then on: every new source anyone finds gets shared to the whole team the day it proves useful. --- ## Stay signed in, stay in control URL: https://dexi.net/kb/account-security-and-sessions Topic: Security Long-lived sessions you can see and revoke per device, Google sign-in, one-click sign-out-everywhere on password reset, and revocable access for every connected app and token. Dexi keeps you signed in for weeks at a time without leaving you blind to where. Settings lists your active sessions — each device with its details and when it was last used — and any of them can be signed out remotely. Signing in works with email and password or with your Google account. Access granted to things that aren’t you is just as visible: connected AI apps (via MCP) appear under Connected apps with a Disconnect button, and personal API tokens are listed and revocable individually. Resetting your password signs out every session everywhere, in one stroke. ### Why it matters A knowledge base concentrates years of your thinking in one place, which makes account control a feature, not paperwork. The principle throughout is: every credential that can touch your account is enumerable and individually revocable — sessions, connected apps, tokens. There’s no “I think I logged in on that old laptop once” anxiety, because you can look. Convenience and control aren’t traded off against each other: sessions last for weeks of normal use so you’re not retyping passwords, but the moment something feels wrong, one password reset ends every session on earth. ### Who gets the most out of it - People who work across devices — Desk machine, laptop, phone — signed in everywhere, visible everywhere, revocable per device. - Consultants on client hardware — Signed in on a client-site machine last month? Check the sessions list and end it from home. - Users adopting AI assistants — Every assistant connection is scoped, listed and one-click revocable — experiment without a leap of faith. - Teams with security review requirements — Clear, user-controlled revocation stories make the tool-approval conversation short. ### Example: The lost-laptop drill Your laptop bag doesn’t come home from a conference. Dexi is signed in on the missing machine. 1. From your phone, you reset your Dexi password. That alone signs out every session on every device — including the laptop, wherever it is. 2. You sign back in and open Settings → Sessions to confirm: the only active session is the phone in your hand. 3. You check Connected apps and API tokens too — the Claude connection and your iOS Shortcut token are yours, still scoped, left alone. 4. Total exposure window: the five minutes it took to notice the bag was gone. Your notes never left your control. --- ## Your first week with Dexi URL: https://dexi.net/kb/your-first-week Topic: Guide A day-by-day plan for going from empty workspace to working system: import, capture channels, tagging habits, a digest, and your first look at the graph. A knowledge system earns its keep through compounding, which means the only real risk is never starting. This is the one-week on-ramp: one small setup step a day, each immediately useful on its own, adding up to the full capture-connect-distill loop by Friday. Nothing here requires discipline or a philosophy of note-taking. It’s plumbing — and once the plumbing is in, the system fills itself. ### Why it matters Front-loading the capture channels matters because capture is where knowledge systems die. If saving something takes more than a few seconds, you’ll stop doing it by week three. Extension, email address, feeds and phone shortcut — once those are wired, the marginal cost of keeping anything drops to near zero. The week ends with the payoff features — digest and graph — pointed at a workspace that already has real material in it, which is the only honest way to evaluate them. ### Who gets the most out of it - Brand-new Dexi users — This is the intended path from signup to working system. - Refugees from other tools — The Chrome import on day one means you start with your history, not a blank page. - Skeptics of note-taking systems — A week of plumbing, no methodology required — judge the compounding on real material. - Team leads rolling Dexi out — Hand this to each new teammate as the standard first-week checklist. ### Example: The seven-day plan One step a day, five minutes each: 1. Monday — capture from the browser. Install the web clipper extension (or bookmarklet) and import your existing Chrome bookmarks. Your workspace starts full, not empty. 2. Tuesday — capture from the inbox. Forward today’s best newsletter to your personal save address (Settings → Save by email) with a #hashtag in the body. That’s the whole email workflow, learned. 3. Wednesday — let sources come to you. Subscribe to three feeds you already trust. Articles now arrive hourly without you patrolling anything. 4. Thursday — start connecting. Write one real note about something you’re working on. Use two #tags and one [[wiki link]], and watch autocomplete suggest tags you’ve already created. 5. Friday — close the loop. Create a digest of your main tag, scheduled for next Friday morning. Future-you now gets a weekly review for free. 6. The weekend — look at what you built. Open the graph view. Even one week in, your bookmarks, notes and feed articles are already clustering around the things you actually care about. Week two has no checklist. The channels are open — just read, save and write like you normally would, and let the system compound. --- ## Earn free notes by inviting friends URL: https://dexi.net/kb/earn-free-notes-with-referrals Topic: Account Every Dexi account has a shareable referral code. When a friend signs up with it, you both get 250 extra free-plan notes — up to 2,500 bonus notes on top of the standard 1,000. The free plan includes 1,000 notes, and referrals are how you grow that without paying. Every account has a permanent referral code — an eight-character code like MAYA7Q2K you can paste as a link (dexi.net/r/MAYA7Q2K) or just read out loud. When someone signs up with it and verifies their email, you both get 250 extra notes. It stacks: each friend who joins adds another 250 to your cap, up to ten friends (+2,500 total — a 3,500-note free plan). Your friends each start with 1,250 notes instead of 1,000, so the invite is a favor in both directions. ### Why it matters The bonus applies to the one limit that matters: because everything in Dexi — bookmarks, emailed notes, feed articles — is stored as a note, extra note capacity is extra room for your whole workspace, not just one content type. There’s nothing to configure and nothing expires. Your code is in Settings → Invite friends, along with copy-link and share buttons and a running count of friends joined and notes earned. Friends can also type the code by hand on the signup page if they heard it rather than clicked it. ### Who gets the most out of it - Students — Class notes, reading and group-project research add up fast — a few classmates joining covers a whole semester without anyone paying. - Heavy free-plan users — Approaching the 1,000-note cap? Two or three invites buys the same headroom as months of pruning. - Study groups and clubs — Everyone who joins through a member’s link starts with a bigger plan, and the sharer’s cap grows with each signup. - Anyone recommending Dexi anyway — If you were going to tell a friend about it, the link version of that recommendation gives you both a bonus. ### Example: A student before the new school year You’ve been keeping class notes in Dexi all summer and your friends are setting up for the fall. In Settings → Invite friends you copy your link and drop it in the group chat: 1. Three friends tap the link, sign up, and verify their email — each one starts with 1,250 free notes instead of 1,000. 2. You get an email as each one lands: your cap is now 1,750 and climbing. 3. A fourth friend hears about it at lunch — no link needed, they type your code into the “Referral code” field on the signup page. 4. By the first week of classes your free plan holds 2,000 notes, and the counter in Settings shows exactly who got you there. The cap tops out at ten referrals (+2,500 notes). Past that, friends who join with your code still get their own 250-note bonus. --- ## Outline as fast as you think URL: https://dexi.net/kb/outline-and-fold-notes Topic: Notes Every list in the editor is an outliner — indent with Tab to nest ideas, then fold branches closed so a long note reads like a table of contents. Long notes have a failure mode: scrolling. The meeting agenda, the project plan, the reading notes that started as ten lines are now two hundred, and finding anything means paging past everything. Dexi builds an outliner into every list — bullet, numbered, or task — so structure and focus are part of the same gesture as writing. Press Tab to indent an item under the one above it, Shift-Tab to bring it back out (the toolbar has matching indent buttons). Any item with sub-items grows a chevron in the left margin — click it to fold that branch closed or open. From the keyboard, ⌘. (Ctrl-. on Windows) folds the item you’re in, and ⌥⌘. folds every branch in the note at once — press it again to expand them all. ### Why it matters Folding turns one note into two views: collapsed, it’s a table of contents you can scan in seconds; expanded, it’s the full detail. That means you can capture at the speed of the conversation — everything indented under the point it belongs to — and never pay for that thoroughness later, because finished branches fold away and your working set stays small. Fold state is remembered per note, and it’s deliberately lightweight: collapsing a branch never marks the note as edited, and undo only ever reverses real changes to your text, not folds. On notes a teammate has shared with you, folding works too — collapse their note however you like to read it, without touching how they saved it. ### Who gets the most out of it - Project leads — A launch plan as a task list with nested subtasks: fold the finished workstreams and the note becomes a live status view. - Meeting-heavy roles — Capture discussion indented under each agenda item, then fold the note down to just the agenda for the recap. - Researchers and students — Reading notes nested by source and theme stay navigable at two hundred lines, not just twenty. - Writers — Draft the argument as a collapsible outline first — reorder and rebalance sections while they’re still one line each. ### Example: A project lead running a launch Your launch note is a task list: five workstreams, each with a handful of subtasks, growing all week as details land in standup. By Thursday it’s too long to scan — so you stop scrolling and start folding: 1. You press ⌥⌘. and the whole note collapses to five lines — one per workstream, each with a chevron and its checkbox. 2. Two workstreams are fully checked off; they stay folded for good. You click open the chevron on “Marketing site” and see the three subtasks still open. 3. In standup you walk the note top to bottom at the folded level, expanding only the branch being discussed. 4. New tasks from the discussion go straight in: type at the end of a branch, Tab to indent under the right workstream, keep moving. Friday’s status update is the same note, folded — you never wrote a summary, you just closed the details. --- ## Study smarter in college URL: https://dexi.net/kb/dexi-for-college-students Topic: Guide One workspace for every course — tag notes by class, link concepts across the syllabus, search by meaning the night before the exam, and get a weekly digest that does your review for you. College notes have a short shelf life by accident, not by design. September’s lecture notes are in one app, the seminar readings are PDFs in a downloads folder, the paper outline is in a Google Doc titled “Untitled document (14)” — and by finals week, the material from the first month might as well not exist. Dexi replaces the pile with one workspace: every lecture note, saved reading and draft lands in the same searchable place. The organizing habit is one keystroke. Type #econ301 or #cogpsych in the note itself and it’s filed under that course. Wrap a concept in [[double brackets]] — [[working memory]], [[comparative advantage]], [[the categorical imperative]] — and it links to your note on that idea, creating one the first time you use it. Lists are a real outliner: Tab indents a point under the one above, and an hour of lecture folds down to just its headings when you’re reviewing. ### Why it matters The payoff shows up at exam time. Full-text search covers every note and tag, and semantic search finds notes by meaning — searching “why do markets underprovide public goods” surfaces your notes even if you wrote “free-rider problem, week 4.” That matters most at the cumulative final, when you’re studying in December with the vocabulary you have now, not the vocabulary you wrote with in September. A weekly digest of each course tag lands in your inbox as an automatic review session — spaced repetition you never had to schedule. It compounds even harder on papers. The web clipper saves journal articles and sources straight into the course tag as you research, wiki links turn scattered readings into a connected argument (the graph view literally draws it), and each concept’s page collects every note that references it — your evidence, pre-gathered before you draft. By senior year, four years of notes are one workspace, which is exactly when a thesis needs them. The free plan’s 1,000 notes comfortably covers a year — and referrals stretch it: each classmate who signs up with your code gets you both 250 extra notes. ### Who gets the most out of it - Students in essay-heavy majors — Papers reward exactly what Dexi keeps: readings, lecture notes and your own ideas, linked by concept and searchable by meaning. - Students juggling classes, jobs and activities — Capture takes seconds — a quick note, a bookmarked article, a thought typed on the bus — and the filing happens by itself. - Study groups — Everyone joining through one member’s referral link starts with extra free notes, and shared reading lists travel as feed invites. - Thesis writers and grad-school applicants — A senior thesis or statement of purpose draws on years of coursework — which is only useful if those years are still searchable. ### Example: A psych major with a term paper due The paper is on memory and studying, due in two weeks, and the semester’s notes are already in Dexi because taking them there was no extra work: 1. All semester, lecture notes went down as outlines — claims at the left, evidence Tab-indented underneath — with #cogpsych in the first line and key concepts in brackets: “[[testing effect]] beats re-reading in almost every study; ties into [[spaced repetition]].” 2. Research week: each journal article and review gets saved with the Dexi extension, tagged #cogpsych #paper on the way in — sources land next to the lecture notes instead of scattering across downloads and open tabs. 3. Outlining starts from the [[testing effect]] page, which lists every note that links to it — three lectures, four saved papers, one late-night idea. The evidence is pre-gathered; the outline is mostly arranging it. 4. Halfway through drafting, there’s a half-remembered study about forgetting curves. Semantic search for “why cramming fades so fast” finds it — filed weeks ago under words that share nothing with the query. 5. Sunday morning, the weekly #cogpsych digest arrives by email anyway — so the material stays warm right through the final, not just the paper. Nothing in that semester was a study system to maintain. It was lecture notes with hashtags — and at finals, the whole term is still one search away. --- ## A feed from your workspace to theirs URL: https://dexi.net/kb/share-notes-between-workspaces-with-feeds Topic: Collaboration Publish a tag as a feed, have another Dexi user subscribe to it, and your notes and bookmarks flow into their workspace — a one-way sharing channel between two accounts, no team required. Teams are Dexi’s tool for working closely with a few people — but plenty of sharing is looser than that: a collaborator at another company, a mentor and mentee, a client you’re curating for. Feeds cover that case. You publish a slice of your workspace as a feed; the other person subscribes to it from their own Dexi account; from then on, everything you tag into the slice arrives in their workspace automatically. The wiring takes two minutes. You create a published feed from a tag — say #design-research — which gives you a public page with a “Subscribe via RSS” link. Send that link to the other person, and they add it under Subscriptions in their own workspace, the same way they’d add any blog or newsletter feed. Dexi checks feeds hourly, so a note or bookmark you tag #design-research this morning is in their workspace by lunch. ### Why it matters What makes this different from sending links in chat is what happens on the receiving end. Your shared items land as first-class content in their workspace: they can tag them into their own system, annotate them, wiki-link them to their own notes, and find them later through search — including semantic search — right alongside everything else they know. You share once; they integrate it their way. And the channel is precisely scoped. Only notes and bookmarks carrying the feed’s tag are published — the rest of your workspace stays private, and sharing something is the same gesture as filing it. It’s also one-to-many: the same feed link works for one collaborator or twenty subscribers, and unlike a team there’s no membership to manage and no limit of one shared team — you can publish several feeds on different topics and subscribe to as many other people’s feeds as you like. ### Who gets the most out of it - Collaborators at different organizations — Co-authors, advisors and research partners can’t always share a team — a feed each way keeps both workspaces fed without merging anything. - Mentors and advisors — Tag the articles and notes worth your mentee’s time, and they arrive in the mentee’s own workspace ready to annotate — a reading list that maintains itself. - Consultants curating for clients — A per-client tag becomes a private-feeling briefing feed: everything you flag for them shows up in their Dexi, with your judgment as the filter. - Study partners and reading groups — Each member publishes a feed of their best finds and subscribes to the others’ — everyone keeps their own notes, everyone sees everyone’s sources. ### Example: Two researchers, two companies, one topic Maya and Sam both work on ML evaluation — at different companies, so a shared team is a non-starter. They set up a feed in each direction: 1. Maya creates a published feed from her #ml-eval tag and sends Sam the public link. Sam opens it, copies the “Subscribe via RSS” URL, and adds it under Subscriptions in his own Dexi. He does the same in reverse with his #evals feed. 2. Their normal work now doubles as sharing. Maya bookmarks a benchmark paper and tags it #ml-eval on the way in; she writes a short note on a failure mode she keeps seeing and tags that too. Both appear on her feed — and within the hour, in Sam’s workspace. 3. Sam reads Maya’s failure-mode note in his own Dexi, tags it into his #robustness system, and wiki-links it to his notes on a related bug. It now shows up in his graph, his searches and his weekly digest, connected to things Maya has never seen. 4. A month in, Sam semantic-searches “why do our evals disagree with user reports” — and the top results mix his own notes with three items that arrived from Maya’s feed. Neither of them ever composed an update email. 5. When Maya starts a draft on the topic, anything she’d rather keep private simply doesn’t get the tag — the feed only ever carries what she deliberately files into it. Two tags, two subscriptions, zero maintenance: each researcher’s curation now compounds in both workspaces. --- ## Knowledge base software that fills itself URL: https://dexi.net/kb/knowledge-base-software Topic: Guide Most “knowledge base software” means help-center tools for support teams. If you want a knowledge base for what you know — your reading, research and notes — the checklist is different: effortless capture, real recall, and a way out. Search for knowledge base software and you’ll mostly find help-desk products: public help centers, article templates, approval workflows, ticket deflection. If you run a support team, that’s the right category — and Dexi isn’t in it. But a lot of people typing that search want something else entirely: a knowledge base for what they know. A place where years of reading, research and working notes accumulate into something searchable, instead of evaporating across browser tabs, inboxes and abandoned docs. Dexi is that second kind — personal knowledge base software, for one person or a small team. Every capture channel produces the same thing, a note: pages you save from the browser, newsletters you forward to your personal save address, RSS articles that flow in, and the notes you type. Structure happens inline as you write — #hashtags become tags, @mentions become trackable sources, [[wiki links]] connect ideas — so the base organizes itself without a taxonomy anyone has to maintain. ### Why it matters The way knowledge base software usually dies is that it becomes a second job. Most tools are built around authoring: the base only grows when someone sits down to “write it up”, which means it grows for three enthusiastic weeks and then stops. Dexi inverts that. Capture is one click from wherever you already are — a browser extension, a bookmarklet, the iOS share sheet, a forwarded email — so the knowledge base grows as a side effect of normal reading and working, not as a separate documentation task. The other half of the test is getting knowledge back out. One search — full-text plus semantic — covers everything you’ve ever saved, so you can find last year’s thinking with this year’s vocabulary. The knowledge graph and similar-notes views resurface connections you didn’t file deliberately. Scheduled digests bring old material back on a rhythm. And there’s no lock-in on the far side: export takes everything as ZIP, CSV or markdown, and the MCP server lets Claude and other AI assistants search and add to the base directly. ### Who gets the most out of it - Solo professionals and lifelong learners — A personal knowledge base rewards whoever feeds it daily — and Dexi is built for the individual reader first, not adapted down from enterprise software. - Small teams of two to ten — Per-note sharing and team-wide search give you a shared knowledge base without a wiki to garden — each person’s normal capture becomes the team’s memory. - People with an abandoned wiki behind them — If your last knowledge base died of upkeep, the fix isn’t discipline — it’s software where filing is the same gesture as writing. - AI-assistant users — A knowledge base your assistant can actually use: connect Claude over MCP and it searches, reads and writes your notes with boundaries you set. ### Example: A consultant replacing a dead wiki You’ve tried this before: a wiki with a careful page hierarchy that went stale within a month. This time you start from capture instead of structure: 1. Day one is wiring, not writing. You install the web clipper, add the share-sheet shortcut to your phone, and set your best newsletters to forward to your personal Dexi save address. Total setup: fifteen minutes, no page hierarchy designed. 2. Then you just work. Articles get saved with a #client-name or #pricing-strategy tag on the way in; call notes mention @clients inline; a stray thought becomes a two-line note from your phone. Nothing is ever “filed” — but everything arrives tagged and searchable. 3. Six weeks in, a new engagement raises a question you half-remember reading about. A semantic search for “why usage-based pricing fails in enterprise” surfaces a forwarded newsletter, two saved articles and your own margin note from a previous client — none of which share your query’s wording. 4. You merge the fragments into one briefing note, wiki-link it to the client’s page, and it joins the graph — the knowledge base got its best page yet, and at no point did you sit down to “do documentation”. The wiki failed because it demanded writing time. This one grew out of reading you were doing anyway.