Zero new capture habit — you already email yourself
Links, X posts, repos, half-formed thoughts — email them to your Moreover address or text them to a Telegram bot. They're fetched, decomposed into claims, routed to the right project, and surfaced the moment they contradict, answer, or supersede something you believe.
Check if your handle is free — live.
Rather look first? See a live workspace — no signup →
No apps to open, no clipper to install, no folders to tend.
Email your addresses — or text a Telegram bot. From your phone share sheet, from anywhere.
Links are fetched — X posts via Grok, repos via GitHub, articles readability-cleaned.
Typed idea / task / todo / read and filed to the right project. Address beats AI; AI fills the gaps.
A belief ledger per project. New claims are tested against it — contradictions surface first.
Daily & weekly digests with one-tap feedback that tunes the engine.
Every project keeps a belief ledger — your assumptions, open questions, decisions. New material is classified against it by a three-stage engine whose default answer is NONE, and whose critic kills every generic card.
What survives is rare and specific. When a card does reach you, it names the delta and what to do about it.
See this in a live workspaceThis article reports FMOs averaging 14-day reconciliation cycles — your April positioning note assumed 30+, which weakens the “weeks to minutes” framing in the pitch.
→ Revise the cycle-time claim in the landing copy
Explicit beats inferred: whatever the address specifies wins, AI fills the rest.
ideas@you.moreover.dev capture an idea — project inferred task+launch@you.moreover.dev a task, filed to “launch” read@you.moreover.dev into the read queue commissionsight@you.moreover.dev anything, straight to that project whatever@you.moreover.dev full AI inference — type and project Per-project assumptions, opens, and decisions. It compounds: every card you accept sharpens the next one. Un-copyable by a tool starting cold.
Not just what you saved — what changed in your thinking. Assumptions challenged, questions answered, decisions made. Daily at your 7am, weekly Friday 4pm; never empty, never noisy.
Ask your archive in plain language. With your key it runs a real tool loop — reading your ledger, searching items, following the thread across projects — not a fixed context window.
Email is one way in; a Telegram bot is the other. Text a link or a task with the same directives (due:, project:, tag:, repeat:). Inbound only, free.
Anthropic, xAI, OpenAI — verified on save, encrypted at rest, last-4 only. Your inference, your costs, near-zero platform margin on AI.
Anyone can feed your instance. Their mail is data, never instructions: review-only, never touches your ledger.
Directives put you in control: date: defers, due:/remind: escalate, repeat: resurfaces a task every N days, expires: cleans up after itself.
Full JSON export and self-serve deletion from day one. No training on your content. Ever.
Private by design, not by promise
The honest objection to any tool like this is trust. So the architecture answers it — these aren't marketing lines, they're how it's built.
Every workspace is its own isolated store, addressable only from your authenticated session. There is no shared table your data could leak through — the isolation was independently audited.
Bring your own Anthropic / xAI / OpenAI key — encrypted at rest, shown last-4 only. Your content goes to your model on your dime. No key? A built-in model runs it, still scoped to you.
Your emails and notes are never used to train models — ours or anyone’s. Full stop.
Content that arrives is treated as untrusted end-to-end: it can’t issue instructions, file to a project on its own, or touch your belief ledger. Only mail from you carries directives.
Full JSON export and one-click account deletion from day one — no email to support, no waiting period.
Bearer-token sessions, SSRF-guarded link fetching, DMARC/DKIM/SPF-verified ownership, per-tenant rate limits — from a documented security pass, not an afterthought.
The full stance lives in the privacy & security docs.