Private beta · Onboarding in waves

Content that gets cited by AI,
not just ranked by Google.

AutoInk is one console for the whole loop: discover keywords, draft publication-grade articles, ship them to your stack, then track which AI engines cite you back — every single day.

autoink.io Invite-only beta GitHub · WordPress · Ghost · Notion · Webflow · Framer · Payload 6 AI engines, scanned daily
autoink.io/inbox

Inbox

6 events 3 unread
New citation on ChatGPT · #1
“The Zettelkasten Method, Explained” is now ranked #1 on ChatGPT for 3 queries.
Inkwell
2h
run_0184 completed
2 articles published to git@main · 184 credits · 6m 22s
Inkwell
2h
5 new keywords discovered
Top score: “obsidian vs notion for writers” · 3,600 vol · score 77
Inkwell
3h
run_0180 failed in publish
Hearth · WordPress 401 — credentials expired
Hearth
5d
ChatGPT rate limit hit
Scan for “offline-first note app” skipped — retrying in 1h.
Inkwell
6h
Perplexity now citing Quill
3 articles from Quill surfaced in Perplexity answers today.
Quill
1d
The new discovery layer

Your readers ask ChatGPT now.
Your SEO stack hasn’t noticed.

Every week a larger share of top-of-funnel traffic starts in a chat window — not a search box. If you aren’t in the answer, you aren’t in the funnel.

AutoInk writes for both worlds, and measures both.

63%
of buyers now start purchase research in an AI chat tool, not a search engine.— AutoInk Buyer Survey · Q1 ’26 · n=1,284
4.2×
conversion uplift on inbound arriving from a cited AI answer vs. a ranked organic result.— aggregate across 38 AutoInk projects
0
major SEO tools that measure AI citations in production today.— as of April 2026
01Keyword discovery

Write what your buyers are asking AI — not what ranked on Google in 2019.

AutoInk mines the long-tail of search, competitor citation gaps, and the prompts people actually type into AI tools — then scores each candidate for answer-worthiness.

01Composite score — volume × intent × citation depth × your topical authority.
02Intent labeled — commercial / informational / transactional, per keyword.
03Status pipeline — discovered → analyzed → queued → generating → published.
autoink.io/projects/inkwell/keywords
IInkwell / keywords 142 total 8 queued
Keyword Vol Diff Score Intent Status
best note taking apps for developers 8,100 42
86
commercial published
markdown editor with git sync 2,400 28
82
commercial published
obsidian vs notion for writers 3,600 31
77
commercial queued
zettelkasten method explained 6,200 38
74
informational queued
daily notes template markdown 1,300 19
68
informational generating
offline first note app 880 24
66
commercial analyzed
knowledge graph note taking 1,100 29
56
informational discovered
autoink.io/articles/zettelkasten-method-explained draft · autosaved
IInkwell / guides / zettelkasten-method-explained draft
zettelkasten-method-explained.md

# The Zettelkasten Method, Explained for Modern Writers

Every productive writer eventually runs into the same wall: thousands of notes, but nothing connects. The Zettelkasten method — German for slip-box — changes what a note is, and how ideas find each other over time.

## One idea per note

A zettel is not a meeting minute. It’s a single atomic idea, written in your own words, small enough to link to from many places without carrying irrelevant context along.

The discipline sounds small. In practice it’s the move that turns a note archive into a thinking tool: once every note is one idea, every link is a claim about how ideas relate.

## Linking, not filing

Instead of nesting notes into folders, you give each one a stable identifier and link between them. Over time, the graph itself becomes the map of your thinking — folders fix notes in one spot; links let them appear wherever they’re useful.

- Fleeting notes — raw captures, triaged weekly
- Literature notes — one per source, in your own words
- Permanent notes — atomic ideas that earn a place in the slip-box

## Why the method stuck

Niklas Luhmann published more than 70 books and 400 papers out of a single 90,000-card slip-box. The system didn’t just store his notes; it produced the next essay he wrote — every morning he started by reading a handful of linked cards and letting the argument assemble itself.

A modern Zettelkasten lives in a text editor. The mechanics haven’t changed: one idea, one note, durable links. The only thing that’s new is how fast you can search it.

Keywordzettelkasten…
Words2,140
Reading9 min
Tokens12.4k
Credits92

Quality0.91

Tags
noteszettelkastenproductivity
Internal links
/how-to-organize…/backlinks-in-notes
02Article draft

Writing your editor would sign off on.

Structured outlines, sourced claims, internal links to your own library — tuned from 20+ of your existing posts so the voice reads like yours, not like ChatGPT’s.

01Quality score on every draft (0–1). Route by threshold.
02Reproducible — reroll with the same seed, diff two versions.
03Credit spend shown up front, before you publish.
03Publish

One run. One commit.
Every stack you own.

AutoInk ships where your team already publishes — GitHub, WordPress, Ghost, Notion, Webflow, Framer, or Payload. Two-way sync: drafts flow out, analytics flow back.

01Atomic runs — prompt, model, sources, output, all captured.
02Rollback any run, any time. Revert the commit, re-queue the keyword.
03Index hooks — IndexNow, sitemap, and search-console ping included.
autoink.io/runs/run_0184 success · 6m 22s

run_0184

success Inkwell 2 articles 184 credits
# 2026-04-20 07:02 UTC  run_0184  project=Inkwell  adapter=github
$ autoink run --project inkwell --articles 2
  → discover: 14 candidates · picked "zettelkasten method explained", "daily notes template markdown"
  → draft: 2 articles · quality 0.91 / 0.88 · 12.4k + 8.9k tokens
  → review: auto-approved · quality ≥ 0.80
  → publish: github@inkwell-studio/site · branch main
     commit f3a29e1  (+2 files, 4,280 insertions)
     sitemap.xml updated
     IndexNow pinged (bing, yandex)
  → announce:
     inbox event run_ok
     slack #content
$ run_0184 complete  6m 22s · 184 credits · 2 articles live
GitHub
commit + pr
connected
WordPress
REST API
connected
Ghost
Admin API
connected
Notion
database
connected
Webflow
collections
connected
Framer
CMS pages
connected
Payload
collections
connected
04GEO · Generative Engine Optimization

The first citation tracker
for the answer-engine era.

AutoInk asks ChatGPT, Claude, Perplexity, Gemini, Grok and AI Overviews the questions your buyers actually ask — every 24 hours — and records which of your articles end up in the answer.

autoink.io/projects/inkwell/geo last scan: 2s ago · ● live
IInkwell / GEO 48 / 112 cited +3 today
Article AI Overviews ChatGPT Perplexity Claude Gemini Grok Δ 7d
The Zettelkasten Method, Explained #1 #1 #1 #1 #2 #1 +2
Best Note-Taking Apps for Developers #2 #1 #3 #2 #4 +1
Markdown Editor with Git Sync #1 #2 #3 #4 +1
How to Organize Research Notes #4 #2 #6 0
Obsidian vs Notion for Writers #3 #5 #8 −1
What “Offline-First” Means for Note Apps #5 #4 +1
Plain-Text Productivity System #7 0
05AutoInk vs. the old stack

Your 2021 content stack
doesn’t measure the new traffic.

before

Your current SEO + CMS combo

discoverAhrefs + a spreadsheet. Weekly export, manual pick.manual
draftBrief → freelancer → 10 days → edits.~$400 / art
publishCopy-paste to CMS, format, add links.2 hr / art
track GoogleGSC + Looker dashboard.ok
track AIYou don’t. No tool measures this.blind
cadence2 articles / month, good weeks.slow
after

AutoInk, one console

discoverWeekly, scored, intent-tagged, queued.auto
draftVoice-tuned, cited, quality-scored in ~3 min.~180 cr / art
publishGit commit, Ghost post, Notion row — one click.0 min / art
track GoogleRank + CTR + impressions, per article.ok
track AI6 engines, daily scan, delta alerts in inbox.live
cadence2 articles / run, 3 runs / week. Your call.24 / mo
06Multi-project ops

Run many properties
from one console.

Separate voice, adapter, cadence and budget per project. AutoInk keeps them all running on the calendar you set — and alerts you when any of them drift.

autoink.io/projects 5 active · 1 paused

Projects

6 total
Inkwell
inkwell.studio
Articles
142
Cited
48/112
GEO43%
Stately Docs
stately.so/blog
Articles
89
Cited
31/78
GEO40%
Hearth
hearth.dev
Articles
64
Cited
19/52
GEO37%
Quill
usequill.com
Articles
108
Cited
42/96
GEO44%
Northlab
northlab.co
Articles
31
Cited
6/28
paused21%
+ New project
autoink.io/scheduler next run: 2d

Scheduler

Mon · Wed · Fri 09:00 UTC
April 2026 1,840 / 2,000 credits
MTWTFSS
06
07
08
09
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
01
02
03
Last run
run_0184 · 6m 22s · ok
07Pricing

One console, all six engines,
every plan.

Every plan includes the full pipeline — keyword discovery, drafting, publishing, and daily GEO tracking across all six answer engines. Plans differ in model access, seats and monthly credit allowance. A credit is what AutoInk spends on an article, a keyword scan, or a GEO pass.

Solo
For one site, getting serious about AI search.
$49 / mo
700 credits / mo included
Join waitlist
1 seat · 1 workspace
GEO tracking · all 6 engines · daily
Auto-scan scheduler
DeepSeek · Claude Haiku drafting
GitHub · WordPress · Ghost · Notion
Most teams
Pro
For teams running a real publishing cadence.
$129 / mo
2,000 credits / mo included
Join waitlist
1 seat · unlimited projects
Claude Sonnet for drafting
GEO: daily + backfill + per-prompt diffs
Voice tuning from your corpus
All adapters + custom webhooks
Slack + webhook alerts
Business
For agencies and platforms running many properties.
$399 / mo
7,000 credits / mo included
Join waitlist
5 seats · unlimited projects
Everything in Pro
Roles, audit log, team billing
Priority support
Shared Slack channel
Free$0

3 credits / mo · 1 project · DeepSeek drafts · Google AIO tracking · try the pipeline end-to-end without a card.

EnterpriseCustom

Custom credit allowance · unlimited seats · SSO / SCIM / SOC 2 · white-glove onboarding · shared Slack. Talk to us.

All plans · cancel anytime · export every article as Markdown + JSON annual billing · 2 months free

Start writing for the
answer engines.

AutoInk is in private beta. Tell us about your site and your content stack — we’re onboarding waitlist teams in waves.

// invite-only · reply to the email we send back · git · wordpress · ghost · payload · notion
FAQ

Questions
we hear most.

Still curious? Book 20 minutes with the team at autoink.io.

How do you “scan” the AI engines, exactly?+
Every 24 hours AutoInk runs the prompts your readers actually ask through each engine’s API (ChatGPT, Claude, Gemini, Perplexity, Grok) and AI Overview scrapes. We parse the citations, match them against your content, and record rank, freshness, and surface context — per prompt, per day.
Will AI detectors flag drafts as AI-written?+
Modern detectors are unreliable — and we aren’t trying to hide. AutoInk writes structured, sourced, factually dense pieces — the kind models reward in citation ranking. The quality score on every run tells you exactly where a draft sits.
Can a human review every article before it ships?+
Yes. Every project has an editorial policy — full auto, review-then-publish, or review every draft. Mix policies across projects, or route by quality threshold.
What happens if credentials expire mid-run?+
The run is held, the draft is preserved, and you get an inbox alert (and email). No silent failures, ever. You see exactly where publish broke.
Who owns the content? Any lock-in?+
You do — drafts, final pieces, embeddings, analytics. Bring your own model keys on Publisher and nothing leaves the vendor perimeter you choose. Exports are plain Markdown + JSON.
Can I try it today?+
AutoInk is in private beta right now — we’re onboarding waitlist teams in waves. Join the waitlist and tell us about your site; we’ll email you when your slot opens. Early-access teams get a free trial (one project, up to 4 articles) before any plan kicks in.

Join the AutoInk waitlist

We’re in private beta and onboarding in waves. We’ll email you as soon as a slot opens — usually within a few weeks.

We’ll only use this email to let you know when your slot opens. No newsletter, no sharing.