The Unbundling of the Operator’s Second Brain
Most product launches I scroll past on Product Hunt are irrelevant to anyone running a cross-border store. But every so often a tiny, single-developer utility lands on something that maps almost perfectly onto an operational problem we all have: too many half-formed inputs — supplier quotes, compliance edge cases, ad-angle ideas, a SKU that keeps triggering returns — and no fast, low-friction place to park them before they evaporate. Gravity Notes, a private offline notepad from solo maker Tristan Manchester, is worth twenty minutes of your attention not because you’ll necessarily buy it, but because its design constraints are a masterclass in the capture-versus-organize tradeoff that quietly decides whether your ops team actually documents anything.
What Gravity Notes actually solves
The premise comes straight from Andrej Karpathy’s idea of keeping one long note: you append new thoughts at the top, and periodically scroll down and bump whatever still matters back up. Manchester tried “maybe every other PKM method available” and found the copy-paste ritual of bumping entries tedious, so Gravity makes each entry its own note in a single list. Type, hit return twice, and it’s saved. Swipe a note right and it climbs back to the top of the stack. That’s the whole mental model.
The details matter more than the pitch. It’s $9.99 once per platform, no subscription — Mac, iPhone, Android, and a command-line tool, with more than 30 releases shipped since the February Mac launch. The editor formats Markdown as you type, so a line starting with - [ ] becomes a tickable checklist directly in the list. There’s a gravity CLI on the Mac that lets Claude Code, Codex, or a cron job search, add, and edit notes, with --json output and a --dry-run flag before anything destructive. Changes sync to iPhone through iCloud. Everything exports as Markdown or plain text, and Android backs up daily to a folder you choose. First-launch setup, larger text sizes, and right-to-left layout for Hebrew and Arabic round out the release notes.
The maker’s stated filter is the interesting part: he keeps saying no to folders, tag managers, and anything that would turn this into “just another notes app.” If a feature slows down capture, it doesn’t go in.
Where the launch itself fell apart
There’s a cautionary tale buried in the comments that every DTC operator should read twice. On launch day, Manchester posted that Apple had taken down all his apps while his developer account renewed — a renewal that had, by his account, already processed days earlier. A commenter in Germany reported the app simply didn’t exist on the macOS or iOS App Store. The maker spent launch day apologizing and collecting country reports instead of converting traffic.
If you sell software, subscriptions, or even just run a brand app as a channel, this is your reminder that platform dependency isn’t an abstraction. A single App Store review or account-state hiccup can zero out your top-of-funnel on the one day you’ve spent weeks priming. Amazon sellers know this intimately — more on that below.
How it differs from the tools you’re probably already using
The comparison the reviewer on the page reaches for is Evernote, arguing Gravity wins because Evernote is “an overkill that takes forever before you are ready to write.” That’s the right instinct but the wrong benchmark for our purposes. Here’s how I’d actually map it against the stack a cross-border operator already pays for.
Against Notion: Notion is a database with a text editor bolted on, and its power is exactly its weakness for capture. Every quick thought becomes a decision about which database, which property, which view. Gravity’s answer — one list, swipe to promote — is the anti-Notion. If your ops wiki lives in Notion and your team never updates it, the problem isn’t Notion, it’s that capture costs too much. Gravity is a clean experiment in dropping that cost to near zero.
Against Obsidian: both are local-first and Markdown-native, so the export paths rhyme. But Obsidian rewards graph-building and linking discipline; Gravity explicitly refuses folders and tag managers. Obsidian is for the person who enjoys the gardening. Gravity is for the person who needs to dump a thought between two supplier calls.
Against Apple Notes and Google Keep: these are the real incumbents, and they’re free. Gravity’s honest pitch is speed-to-first-character plus the CLI, not features. If you’re already fast in Apple Notes, the switching case is thin — and I’d say that plainly.
Against Todoist or Things: Gravity is not a task manager and shouldn’t be treated as one. The checklists are for in-note ticking, not for project management with due dates and assignees. Anyone who tries to run a launch calendar in it will be disappointed, correctly.
Why Amazon sellers should care more than Shopify ones
This is the sidebar I’d underline. A Shopify operator’s institutional knowledge tends to live in a handful of places: the store’s admin, a Klaviyo account, a spreadsheet, and maybe a Slack channel. Messy, but centralized-ish.
An Amazon Seller Central operator is the opposite. Your reality is fragmented across listing health, Helium 10 or Jungle Scout research exports, case logs, A-to-z claims, reimbursement tickets, and the running folklore of “that SKU always gets flagged in Q4.” None of that fits neatly into a database schema, and most of it never gets written down at all. A frictionless, offline, plain-text capture list — one that an AI agent can read via CLI and summarize into a case file — is genuinely more useful to an Amazon brand owner than to a Shopify merchant, because the Amazon operator’s biggest undocumented asset is accumulated pattern recognition about the marketplace itself.
The offline, private framing also matters here. Supplier terms, margin structures, and account-health war stories are not things you want living in a free tier of somebody else’s cloud.
The CLI is the actual headline
The swipe gesture is charming, but the gravity command is why this belongs in a piece about operator tooling. The maker built it because he wanted to jot something down and have his OpenClaw agent ask him about it in the coming days. That’s a workflow pattern, not a feature: capture in plain text, let an agent query the corpus later, get synthesis back.
For a cross-border seller, translate that to: dump every supplier email snippet, return reason, and ad-angle hunch into one list over a quarter, then point Claude Code at it with --json output and ask it to cluster the return reasons by SKU. That’s a poor man’s voice-of-customer engine built on a $9.99 notepad instead of a five-figure analytics contract. The --dry-run flag before destructive edits is exactly the kind of guardrail you want when an agent has write access to your notes.
What cross-border sellers can borrow from the design
Even if you never install it, Gravity’s constraints are a template worth stealing for your own internal tooling and SOPs.
Capture must cost less than the thought is worth. Manchester’s rule — if a feature slows down capture, it doesn’t go in — is the correct priority order for any ops documentation system. Most teams build the organize layer first and wonder why nobody feeds it. Invert it. Make the inbox stupidly fast, then add structure only where retrieval actually fails.
One list beats many folders for early-stage signal. Folders force premature categorization. A single chronological stack with a promote-to-top gesture is a natural ranking mechanism: what you keep bumping is what matters, and the ranking emerges from behavior rather than from a taxonomy meeting. For product research, this is a useful way to let winning SKU candidates surface themselves — the ones you keep returning to are the ones worth sampling.
Plain-text export is a strategic asset, not a checkbox. Markdown and plain-text export means your notes survive any tool’s death, and they’re directly ingestible by LLMs. If your current knowledge base locks content in a proprietary format, you’re one pricing change away from losing your institutional memory — or from being unable to feed it to an agent.
Agents need write access with guardrails, not read-only dashboards. The --dry-run pattern is the right compromise. Give your automation the ability to propose edits to your ops notes, review the diff, then commit. That’s how you get the leverage of automation without the 3 a.m. panic of an agent “helpfully” rewriting your supplier terms.
Where the math breaks
I’ll be blunt about the limits. Single-developer software is a real risk for anything you’d build a workflow on. The launch-day App Store takedown is a live demonstration that the distribution layer can vanish for reasons outside the maker’s control, and the maker’s own comment thread shows him fielding country-availability complaints and offering promo codes by email — heroic, but not a support organization.
The feature ceiling is also real by design. No multi-word tags yet, no tag autocomplete (the sole reviewer’s two improvement notes), no folders, no team collaboration, no shared workspaces. For a solo operator that’s fine. For a five-person ops team that needs a shared, permissioned knowledge base, this is not the tool — and the maker would presumably tell you so.
And the $9.99-per-platform pricing, while refreshingly subscription-free, means you pay separately across Mac, iPhone, and Android. For a team spanning a warehouse in Shenzhen, an ops desk in Manila, and a founder in Berlin, that’s three purchases per person and no shared list. The economics favor individuals, not teams.
What I’d watch / test next
Three concrete moves for this week, in order of effort.
First, steal the pattern before you steal the product. Open a single plain-text or Markdown file — one list, newest on top — and for seven days dump every operational observation into it without categorizing. At the end of the week, count how many entries you’d have lost otherwise. That number tells you whether capture friction is actually costing you money.
Second, if you run an Amazon brand, test the agent loop cheaply. Export a month of return reasons and case notes into Markdown, point your LLM tool of choice at the folder, and ask it to cluster by SKU and root cause. No new subscription required — you’re validating whether synthesis-on-demand beats a dashboard you never open.
Third, put a platform-dependency audit on your calendar. List every channel where a single account state — App Store review, Seller Central health, a payment processor hold — could zero your funnel overnight, and write the contingency line for each. Gravity Notes’ launch day is the cautionary tale; your job is to make sure it stays somebody else’s story.
I’d also watch whether the CLI gets a proper team story — shared lists, conflict handling, permissions. That’s the feature that would move this from a clever solo tool to something an ops department could standardize on. Until then, it’s a sharp little instrument with a clear philosophy, and the philosophy is the part worth importing.
