DeskMemory on Training Wheels: The Features Nobody Tells You About (Until Now)

# DeskMemory on Training Wheels: The Features Nobody Tells You About (Until Now)

You’ve started your vault. You’ve asked the AI a question, felt that little spark when it answered from your own material. Now it’s time to stop being a passenger and start driving.

Most new users discover the obvious stuff — quick capture, chat, search. But DeskMemory hides a whole layer of true mastery just below the surface. These aren’t “advanced” features in the sense of complexity; they’re “advanced” in the sense of control. Using them feels like unlocking a secret menu.

Let’s walk through the ones that turn a cluttered folder into a personal AI command centre — with zero coding, zero markdown expertise, and plenty of “oh, that’s neat” moments.

1. The Quick Open Palette: The Fastest Route to Any Note

You know the drill: you want to get to a note, you scan the sidebar, you scroll… no, wait, you search, you scroll more. Not anymore.

Hit Ctrl+P (or ⌘P on Mac) and a floating palette appears. It’s your teleport.

Here’s what it can do right now: - Show recent files — no typing needed, the top of the list remembers where you’ve been. Expand the “Recent files” chevron to browse up to a handful of paths. - Filter by file name — just like the sidebar filter, but faster. Type `proposal`, and any note with that word in its name lights up. - Go full meaning-based — start with `~` and the palette switches to semantic search. `~ client onboarding checklist` finds notes that are conceptually about that process, even if the file name says something else. - Hybrid mode — if you just type words *without* `~`, DeskMemory blends name matches with a few top semantic suggestions, marked clearly in the results. Best of both worlds, no special syntax needed.

Training habit: Use Quick Open at least three times today instead of clicking through the tree. By Friday, your hand will go to Ctrl+P automatically — and you’ll save minutes every time you sit down to work.

Wow moment: A colleague asks for that old campaign analysis you did months ago. You type `~ campaign analysis 2025`, spot the note in the palette, hit Enter — five seconds, conversation saved.

2. “I pasted a file and now it’s a mess” — Meet Edit Details

You drag in a `.md` from another tool. Or you paste a rich text snippet from the web. The result: a note with no category, no tags, no proper title, and a cryptic filename like `download-12.md`. Until you tame it, the AI can’t classify it properly, and you can’t browse by category.

Rather than learning YAML front matter (those dashed lines at the top of a note), just use the Edit details… button that appears in the editor header.

• Click it. A pop‑up asks for title, type (note, master, etc.), category, date, and tags.
• Fill in what you know. If you leave the date blank, it defaults to today — no pressure.
• DeskMemory then writes (or updates) the tiny block at the top of the file, renames the file from the title, and moves it into its category folder (creating the folder if needed). Your note body stays untouched.
• The sidebar immediately updates. The AI now knows where it belongs.

Training exercise: Open any unclassified note from your “root” folder right now — the one with the weird name. Use Edit details to give it a title and category. Do one more tomorrow. In a week, your vault will feel organised without a single manual drag‑and‑drop.

Wow moment: That pile of unsorted imports? After three sessions with Edit Details, your library suddenly has folders, tags, and a logic the AI genuinely uses.

3. Control What Search Sees (Without Deleting)

Not every note deserves to appear when you type `~`. Maybe it’s a template, a previous version, or your half‑baked venting from a bad day. You want it in the vault, but you don’t want it muddying your results.

On any note that already has front matter, the editor header shows a toggle: Include in vault search.

• Turn it off for reference files, old drafts, or personal diaries. The sidebar marks them with a little eye‑off icon.
• Turn it on for anything you want the AI and semantic search to treat as “source material.”
• Current master notes (like your brand guide or résumé) always stay searchable — the app knows they’re too important to hide.

Training habit: After a big capture session (a dozen recipes, ten meeting notes), spend 30 seconds reviewing the list. Flick “Include in vault search” off for the raw‑imported versions, and keep it on for the polished ones you’d actually want the AI to reference.

Wow moment: Your semantic searches suddenly feel sharper — because the noise is gone, and the signal remains.

4. Build Your AI Toolbox (Taxonomy That Works Hard)

This is the quietest superpower in DeskMemory, and it’s pure “set it once, benefit forever.”

Under Settings → Chat Settings, you can define:

• Categories — `Recipes`, `Clients`, `Travel`, `Operations`… these are the buckets that show up in the Save dialog and in your tree’s folder structure.
• Tags — `pasta`, `client‑x`, `budget`, `draft` — flexible labels that the AI can later use to pre‑fill when you save a chat (with AI prefill enabled).
• Role presets — e.g., “strategic consultant,” “private chef,” “HR policy writer.” These are fed to the AI as tone instructions when you create prompts and when you use a template.
• Output format presets — “email,” “report,” “blog post,” “itinerary” — these shape how the AI structures its answers and how the Prompt Compiler packages them.

Training exercise: This week, add three categories and five tags that reflect your real work. Don’t overthink — you can edit them later. Once they’re in, every time you save a note, you’ll see chips for those tags. The AI will also use this taxonomy to pre‑fill classifications (when AI prefill is on). You’re basically teaching DeskMemory your vocabulary.

Wow moment: A month from now, you type “`~ pasta`” and every recipe is already there, pre‑tagged. The AI proposes tags like `vegetarian` automatically. It feels like the app grew along with you.

5. Chat Like a Pro: Multiple Threads, No Headaches

DeskMemory’s chat isn’t a single blank page — it’s a workspace with multiple simultaneous conversations, each one saved automatically and ready to pick up where you left off.

• The Open chats rail (left side, collapsible) holds every session. The active chat sits at the top; click any other to switch instantly.
• Each chat remembers its own settings — tone, web search preference, attached vault files, even the unmade draft you were typing. Close the app, reopen it, everything restores.
• Working on a proposal? Start that chat tab. While you wait for the AI to draft, switch to another tab and ask a quick question about recipes. Both streams keep running.
• When a chat is done, hit the × to close it. DeskMemory will nudge you to Save to vault first — giving that thread a permanent home.

Training habit: Create a new chat for each project or topic. Name the first user message clearly (“Research for Q3 campaign,” “Draft apology email for guest”). That label becomes the chat’s title in the rail, making it easy to pick up later.

Wow moment: You never again lose the brilliant thought you typed three chats ago — because every chat persists, and the composer drafts survive even a restart.

6. See What the AI Sees (Transparency Mode)

DeskMemory treats you like a co‑pilot, not a passenger. Two features give you a window into the AI’s thinking:

Rich replies vs raw Markdown In any chat, you can toggle between Formatted view (clean, rendered responses) and Raw Markdown view (exactly what the model returned). Use this when you want to copy the source formatting, inspect a link target, or just satisfy your curiosity. It’s also a gentle way to learn Markdown without pressure — you see the `bold` and `## headings` as you read.

The Reasoning block (Research mode) When you use the Research (deeper) AI mode, the model’s intermediate reasoning appears in a collapsible block above the final answer. While streaming, it opens automatically; you can fold it away or copy it along with the answer. This is brilliant when you’re making a serious decision — you can see the “how” as well as the “what.”

Training tip: Switch a chat to “Research” and ask a question that has no single right answer: “Compare two vacation destinations.” Watch the reasoning unfold, then check the sources when web search is on. You’ll start to trust the AI more — and you’ll spot when it’s interpreting, not just regurgitating.

Wow moment: You’re not just getting an answer; you’re getting a glimpse of the process — and that makes you smarter about your own thinking.

The Takeaway

Mastering DeskMemory isn’t about memorising commands; it’s about adopting a few small habits that compound over time. Use Quick Open to teleport. Tame your imports with Edit Details. Clean your search scope. Build your personal taxonomy. Treat chat tabs like workspaces. And peek under the AI’s hood now and then.

Each one feels tiny in the moment. But stack them together, and you stop being a user — you become the commander of your own knowledge.

Now go try one. Just one. Today.

*Start my free library → link*

!Deskmemory3