Forward content in.
Get a trustworthy call out.
Intrinsic Cognition is an AI signal-triage tool. You forward any link, post, doc, or email. It fact-checks it, validates the source, and tells you whether it’s worth your time now, later, or not at all — with a clear reason every time.
Free during the beta. No credit card required.
What it does
Three honest answers for everything you forward.
For each item, you get the central claims fact-checked against real sources, an assessment of substance and recency, and one of three calls — with a rationale you can disagree with.
Use now
High signal for an active project. Comes with the project it ties to and a concrete suggested next action.
Defer
Right idea, wrong moment. Gets parked against the phase it will matter in, then resurfaces automatically when that phase arrives.
Ignore
Filtered out — with the reason: irrelevant, low quality, already known, or superseded. Stays visible (collapsed) so the filter is auditable.
How it works
Capture is instant. Triage is async. Surfacing is quiet.
One loop, six stages. Built to protect your attention, not compete for it.
- 01
Forward it
Share to a dedicated email address, paste into the app, or use the iOS share-sheet shortcut. Capture is instant — you move on.
- 02
It gets normalized
Links, posts, screenshots, emails — extracted to a canonical text form, with provenance (URL, author, platform, publish date) preserved.
- 03
It gets fact-checked
The central claims are pulled out and verified against the open web. You see per-claim verdicts with sources, plus an honest "unverifiable" where the evidence is not there.
- 04
It gets a call
Validated against your active projects and their current phases: use-now, defer-to-phase-X, or ignore — always with rationale you can override.
- 05
It lands in your digest
New use-now items at the top, a scannable list of triaged items, and ignored items collapsed but visible — so the filter stays auditable.
- 06
Deferred items come back
When a project enters the phase you deferred for, the item resurfaces — framed as "you deferred this then, specifically for this phase." A defer that never comes back is just a polite delete.
Design principles
Trust is the product.
A triage tool you can’t inspect is worse than no triage at all. These are the constraints we build under.
Fact-checks must show their work
Every checkable claim is verified against real, linked sources. The system is allowed — expected — to say "unverifiable" rather than guess. A fact-check with no source isn't one.
Defer must really come back
Items you defer are tied to a project phase and resurface when that phase arrives. The whole tool only works if defer is honest, so the resurfacing is a first-class feature, not a nice-to-have.
Ignored items stay inspectable
Filters earn trust by being auditable. Ignored signals are collapsed in your digest, not deleted, so you can catch the filter making mistakes and correct it.
Forwarded content is data, never instructions
A forwarded post that says "ignore previous instructions" is treated as text to triage, not a command to follow. The triage agent runs least-privilege by design.
Quiet by default
Only high-confidence use-now items interrupt you. Everything else waits for the digest. We don't notify to drive engagement; we notify when there's genuinely something to say.
Your corrections matter
One click overrides a decision or flags a bad fact-check. Corrections are logged with the signal so the filter sharpens to your judgment over time — and so the tool's accuracy stays auditable.
Request beta access
Onboarding is one cohort at a time.
We’re starting with friends, family, and a small private group so we can tune the triage before opening it up. Drop your email and we’ll be in touch when a spot opens, with what to expect before you say yes.