Features Group II — Intelligence · Element 18

Smart Archive —
evergreen notes,
on autopilot.

The notes you keep coming back to deserve a permanent place. The ones you don't deserve a graceful exit. Knovya watches the maturity signals — backlinks, NoteRank, edit cadence, age — and graduates the right ones for you. Andy Matuschak's philosophy. The system's discipline.

5
lifecycle
stages
5
outcome
categories
0
manual
maintenance

§ 2 — The Library

Working notes. Candidates. Evergreen.

Three columns, one workspace, one weekly graduation pass. The notes on the left are still in motion. The notes in the middle have stopped moving — they earned the right to be considered. The notes on the right have graduated. Hit the run button and watch the candidates make their move.

Stage 01

Working 4

Active drafts and in-progress thinking. Edits are still landing.

draft

Q4 capacity model — early sketch

"Headcount-aware ship velocity, working assumptions only."

fresh0 backlinks
meeting

Customer call · Acme Co · this Tue

"They asked about predictable releases."

this weekno edits yet
draft

Pricing tier rewrite v3 — outline

"Three tiers, hybrid usage cap, working through edge cases."

edited 4x1 backlink
plan

Hiring plan — open senior IC slots

"Three slots open, capacity-gap-prioritized."

activeedited 6x2 backlinks
Stage 02

Candidate 3

Completed and well-cited. Maturity signals stacked. Ready for the weekly check.

plan

Q3 OKRs — leadership offsite

"Cross-team OKRs, finalized after the offsite."

completededited 14x8 backlinksranked high
decision

Pricing tier review — April spike

"Three flat tiers, no usage caps. Tested for one cycle."

completededited 11x5 backlinks60 days old
research

Senior IC retention study — 2024 cohort

"Stalled mid-tenure work weighted heavier than comp."

completedcited 7xranked high
Stage 03

Evergreen 3

Graduated. Carries an outcome and an agent badge — your long-term reference shelf.

decision

Auth security review · April

"7-day token rotation policy, replaces 30-day plan."

linked from 4 downstream
success

well-cited and consistently surfaced; load-bearing for the auth flow

plan

Onboarding rewrite — v2

"Three-step flow, voice-first option, OAuth retry path."

linked from 6 downstream
partial

matured over 90 days; some hypotheses landed, OAuth retry partial

decision

Dual-pricing experiment · postmortem

"Both anchors hurt conversion. Documented as cautionary."

cited 9x as warning
cautionary

kept precisely because it failed; future plans on this shape get flagged

§ 3 — The anatomy

Five lifecycle stages. Five archive outcomes.

Every note in Knovya has a lifecycle. Most notes never reach the last stage — that's the point. The ones that do, do so for reasons the workspace can name. The ones that get archived earn an outcome that becomes precedent for future work.

3a · Lifecycle stages

A note's path from draft to long-term reference.

Draft

The early shape — first sketches, partial structure.

Active

The working phase — being edited, linked, used.

In-progress

The iteration phase — checklists in motion, decisions firming up.

Completed

The closure phase — the work is done, edits stabilized.

Evergreen

The long-term phase — load-bearing for future work, kept active.

3b · Outcomes when archived

The five categories that turn an archive into precedent.

success

Success

It worked. Stamp it as positive precedent — Experience Envelope will surface this as encouragement next time you face the same shape of problem.

partial

Partial

It half-worked. Some hypotheses landed, some didn't. Stays useful as nuanced reference rather than clean evidence either way.

cautionary

Cautionary

It failed. Worth keeping precisely because it failed — future plans on the same shape get flagged with the lesson.

cancelled

Cancelled

It was abandoned before resolution. Not a failure — just dropped. Kept for traceability when someone asks "what happened to that?"

superseded

Superseded

It was replaced by a later version. The supersedes chain points to the successor — agentic memory walks it during temporal recall.

§ 4 — The problem

Most notes never graduate. They die in your inbox.

Scene 01

The manual evergreen practice — Andy Matuschak's five years, Maggie Appleton's seedling-to-evergreen garden — is admirable and effective. It's also a discipline almost nobody sustains for a quarter, let alone half a decade. The framing is right; the friction wins.

Scene 02

Notion, Obsidian, Roam, Logseq give you the buttons — archive, pin, tag with "evergreen." None of them carry an opinion about when. Every user reinvents their threshold every Monday, and most users abandon the practice within a quarter.

Scene 03

Apple Notes, Bear, Day One ship the archive button and stop there. No maturity intelligence, no outcome enrichment, no precedent record. Archiving means "I'm done with this" and nothing more.

The signals that distinguish a working note from a long-lived one are knowable — backlinks accumulate, NoteRank rises, edit cadence stabilizes, age compounds. Knovya watches the signals so the discipline doesn't have to be yours. The philosophy stays Andy's. The discipline becomes the system's.

§ 5 — Lineage

Sixty years of permanent notes.

The idea that some notes are meant to last is older than the personal computer. Each milestone in the lineage is the same insight — that knowledge work needs a long-term reference layer — implemented with the technology of its time. Smart Archive is the first version where the curation runs on its own.

  1. 1960s—1990s

    Niklas Luhmann's slip-box

    German sociologist Niklas Luhmann ran a Zettelkasten of 90,000 paper index cards, each one an atomic permanent note linked to others by hand. The first systematic evergreen practice; the foundation everyone after him cites.

  2. 2017

    Ahrens — How to Take Smart Notes

    Sönke Ahrens' book formalizes Luhmann's method for the digital era — fleeting notes, literature notes, permanent notes — and makes "permanent notes" a term the broader knowledge-work audience knows.

  3. 2019—

    Andy Matuschak — Evergreen Notes

    notes.andymatuschak.org publishes a working notes site as a public artifact. Andy's evergreen-notes framing — atomic, declarative, statement-titled, evolving across projects — becomes the canonical reference for the modern PKM movement.

  4. 2020—

    Maggie Appleton — Digital Garden

    Maggie Appleton's digital garden establishes the seedling → budding → evergreen growth-stage metaphor and visualizes the maturity arc most note-takers feel intuitively but rarely encode.

  5. 2026

    Knovya — Smart Archive

    First automated evergreen curation. The maturity signals — backlinks, NoteRank, edit cadence, age — feed a daily graduation engine. Andy's philosophy. Maggie's growth stages. The system's discipline.

§ 6 — First mover

Manual evergreen curation existed for sixty years. Automated evergreen curation didn't exist at all. Until now.

Andy Matuschak · Maggie Appleton · the manual practice

A discipline practiced by people who can sustain it. Five years of evergreen-note maintenance is admirable, rare, and not within reach for most knowledge workers managing a normal job alongside their note-taking.

Notion · Obsidian · Roam · Logseq

Buttons to archive, pin, tag — but no opinion about when, and no signal aggregation. Every user reinvents the threshold. Most users abandon the practice within a quarter; the buttons stay, the curation doesn't.

Apple Notes · Bear · Day One

Archive UI exists; maturity intelligence does not. Archiving means "done with it" and nothing else — no outcome stamp, no precedent record, no link back to future work.

Knovya — Smart Archive

Automated graduation from maturity signals plus outcome-stamped archive. The signals that distinguish a long-lived note from a transient one — watched daily, surfaced weekly, applied with your consent. The Andy Matuschak philosophy without the Andy Matuschak workload.

§ 7 — Surfaces

Where Smart Archive surfaces.

The library runs in the background. The surfaces are the touchpoints where you decide — daily suggestions, pre-archive impact previews, precedent banners on similar work, and the evergreen library itself.

Surface 01 · Daily inbox

Today's graduation queue

Every morning, Smart Archive surfaces the notes whose maturity signals stacked overnight. You see three actions, one click each — graduate to evergreen, archive with an outcome, deprecate with a reason. Inbox-shaped, not settings-shaped.

Today's queue 3 candidates
graduate Q3 OKRs — leadership offsite plan accept →
archive Pricing tier review · April spike review →
deprecate Old auth flow — pre-OAuth replace →

Surface 02 · Pre-archive

Impact preview before you commit

Archiving isn't a one-click destruction — it's a lifecycle transition. Knovya shows the impact (dependent notes, backlinks, active shares) and asks for the outcome stamp before the move. The outcome becomes precedent, not just metadata.

Archive: "Pricing tier review · April spike"

Before this archives, here's what's connected:

5 backlinks from active notes — Q3 OKRs, hiring plan, roadmap
2 downstream notes depend on this — heads-up sent
1 active share — collaborator notified

Pick the outcome — feeds Experience Envelope:

success partial cautionary cancelled superseded

Surface 03 · Precedent

When the past speaks up

When you start a new note that looks like one you've shipped before, a precedent banner surfaces the past outcome — Experience Envelope's read of which similar notes succeeded, partially landed, or went cautionary. Past work, applied to present work.

New note: Q4 capacity model

Headcount-aware ship velocity for Q4 planning, with comparisons to Q3 baselines and a note on hiring assumptions.

Precedent · 4 similar notes archived

This kind of capacity model has shipped 3 times with success and once with a partial outcome. The partial archive flagged "headcount assumed without hiring lag" — worth checking.

Surface 04 · The library

Filter by outcome — find the success patterns

The evergreen library is a real workspace view. Filter by outcome to surface the success archive (what worked), the cautionary archive (what didn't), or the partial archive (what nuanced). The shelf you actually go back to.

all evergreen success partial cautionary
Auth security review · April success
Q3 OKRs — leadership offsite success
Onboarding rewrite · v2 partial
Dual-pricing experiment cautionary

§ 8 — Bonded with

Smart Archive doesn't graduate alone.

The maturity signals come from somewhere. NoteRank weighs which notes are load-bearing. Experience Envelope reads the outcome stamps as precedent. Reflect's Crystals graduate naturally to evergreen. Knowledge Graph supplies the link density that distinguishes a working note from a long-lived one.

Sa · 18 · ★

The discipline you'd never sustain.

Andy Matuschak ran his evergreen practice for five years. Maggie Appleton has tended hers for six. You won't. That's not a personal failure — that's how friction works. Knovya watches the signals so you don't have to remember to.

Frequently asked.

What are evergreen notes?

An evergreen note is a long-lived, declarative, atomic note that earns its place in your knowledge base because you keep coming back to it. The term was popularized by Andy Matuschak's working notes — short, statement-titled notes that evolve, contribute, and accumulate across projects rather than getting archived after one use. Maggie Appleton's digital-garden growth-stage metaphor calls them the final stage after seedling and budding. The practice predates the name: Niklas Luhmann's Zettelkasten slip-box ran on the same idea with index cards.

How does Smart Archive decide which notes are ready?

Knovya watches the signals that distinguish a working note from a long-lived one — how often it's edited, how many other notes link to it, whether NoteRank consistently surfaces it, how old it is, whether it's reached a closure status. When the signals stack up, the note becomes a graduation candidate; the daily archive-suggestions surface it for one-click acceptance, or the workspace can apply graduations automatically based on your settings.

Can I manually promote a note to evergreen?

Yes. Every lifecycle transition has a manual control — promote, archive, deprecate, restore — and your manual choice always overrides the system. The Smart Archive runs as a daily suggestion engine; you stay in charge of what graduates and what gets archived. Manual promotions skip the candidate phase entirely.

What's the difference between archived and evergreen?

Archived means the work is done and the note is being kept for reference — it leaves the active workspace but stays searchable. Evergreen means the work is done AND the note is load-bearing for future work — it stays in the active workspace as a long-term reference. A failed experiment gets archived with a cautionary outcome; a successful pattern gets graduated to evergreen. Both keep their backlinks and supersedes chains intact.

Can I see why a note was promoted?

Yes. Every graduation comes with a short qualitative reason — "well-cited and consistently surfaced," "matured over 60 days, edits stabilized," "linked from four downstream decisions." The reasons are natural-language summaries of the maturity signals; the underlying weights and thresholds stay internal because they're an implementation detail you don't need to argue with.

Does Smart Archive work for new users?

Smart Archive needs a few weeks of activity before its signal is reliable — there's no useful maturity score for a workspace that's three days old. Until then, you can promote notes manually; the daily suggestions become meaningful around the time the workspace itself does. Cold-start workspaces see manual lifecycle controls only; the suggestion engine activates as the signal accumulates.

Can I disable auto-promotion?

Yes — workspace settings carry three knobs. You can leave Smart Archive in suggest-only mode (the system flags candidates, you accept manually), turn on auto-graduation for high-confidence cases (the system promotes, you can revert), or scope the whole feature to specific folders so a personal-projects folder stays manual while your work folder runs on autopilot.