Bring a paper you believe contains a mechanism for your problem.
Upload a paper, framework, or source idea; choose the region you want to project it toward; and receive hidden mechanisms, multidisciplinary bridges, fit labels, a Trajectory Tree, and next verification steps a single-field reader may miss.
The method, plainly: we extract the source's underlying way of doing things, transplant it into the problem space you choose, and hand you the verification step to run — a mechanism, moved to where it may quietly help.
And we are honest about fit: sometimes the mechanism transplants into your niche, sometimes it does not. Every projection is labeled real bridge, speculative, or no match — with the reason. A clear "this does not fit, and why" is part of what you pay for.
The deliverable is an insight card.
Each card is a candidate research direction with mechanism, bridge, evidence, risk, and verification step.
Inner insight
What did the source really discover underneath the surface vocabulary?
Problem projection
Choose biology, economics, control theory, neuroscience, law, business, or another problem space and ask what the mechanism becomes there.
Trajectory Tree
Included in every report: the path the field actually took, and the latent branches it left behind — each with a first move. See an example ->
Your intuition aims it. The engine carries it.
Picture every paper or source idea as a point in one vast space. The engine extracts its way of doing things, but your intuition chooses the problem region worth aiming at. That chosen region is where the method is tested honestly.
When a mechanism does not fit your narrow space, that is a signal, not a dead end: widening the field — a nearby subfield, an adjacent problem — is often exactly where the transplant starts to work.
What the cards feel like.
This is a real run. We ingested AlphaFold — one of the most-cited papers of the decade — and asked Ingest Paper to project its hidden mechanism into other fields. Three of the leads it surfaced:
Folding -> Economics
What we saw: AlphaFold's "constraint-satisfaction toward a single low-energy state" is the same shape as a market settling into equilibrium under competing forces.
Bridge: Treat price discovery as a folding problem — many local pressures resolving into one stable configuration.
Folding -> Quantum Physics
What we saw: The model collapses a vast space of possible structures onto the one nature selects — echoing how measurement collapses a superposition.
Bridge: Read structure prediction as a selection rule over a state space, not a search.
Folding -> Game Theory
What we saw: Each residue "plays" against the others; the folded protein is the Nash equilibrium of the chain.
Bridge: Frame folding as a multi-player game whose stable outcome is the native structure.
Open the full AlphaFold report -> · Open the full Transformer report -> · Open the full Leibniz Monadology report ->
Five steps, no magic.
The engine does the heavy lifting on either side — and in the middle it even scouts the likeliest fit. But the aim is also yours to take.
Start with one directed projection, or explore wider.
We accept up to 5 new paid reports per week, and Founding Beta pricing applies to the first 50 paid reports. Every report is generated with AI systems and manually reviewed before delivery. Directed projections count inside the promised card total; unused directed slots become scouted cards. Projections beyond an offer's included allowance add one card for $25 each.
Try One Projection
- 1 directed projection card
- Hidden mechanism and honest fit verdict
- Mapping, risk, and verification step
Explorer
- Up to 2 directed projection cards
- Enough scouted cards to reach 5 total
- Compact Trajectory Tree
Research
- Up to 3 directed projection cards
- Enough scouted cards to reach 10 total
- Full Trajectory Tree
- 1 Depth-2 credit and founder review
Frontier
- Up to 5 directed projection cards
- Enough scouted cards to reach 25 total
- Full Trajectory Tree
- 2 Depth-2 credits, founder review, and a short founder note
Golden Garden: not a tier, a project.
Reports explore. Deepen focuses. When a lead becomes a real, defensible hard problem, the path moves to Golden Garden: founder-led work on one high-stakes problem, with room for protectable or licensable value.
- Opened by qualification, not impulse checkout.
- Triggered when a lead clears our private Golden Garden screen.
- Starts with a guided discussion and founder review.
- Commercial structure can include engagement fee plus shared upside.
All prices are USD. Additional directed projections beyond the included allowance are $25 each and add one card each. Follow-ups: Depth 2 $35, Depth 3 $60, or three Depth-2 credits for $79. The first eligible widening is free; later widening is $15. Secure checkout is handled by Stripe. No card data is collected on this website. Delivery happens only after founder review.
Do directed projections replace scouted cards?
They count inside your promised total. Explorer with two directed targets delivers those two directed cards plus three scouted cards. If you use fewer than the allowance, the unused slots become scouted cards.
What happens if I choose more directed targets than my offer includes?
Each additional target costs $25 and adds one card. For example, Explorer with three directed targets is $104 and delivers six total cards: three directed plus three scouted.
What does Try One Projection include?
One required directed projection with the hidden mechanism, fit verdict, mapping, risk, and next verification step. It is the smallest paid test of a specific intuition.
What if the result is “no match”?
An honest, supported no-match is a valid result. For an eligible no-match, the first widening is free; later widening costs $15. If our quality or safety review blocks delivery, we offer a corrected rerun or full refund.
How do follow-up credits work?
Research includes one Depth-2 credit and Frontier includes two. After credits are used, Depth 2 is $35, Depth 3 is $60, and a three-credit Depth-2 pack is $79.
When will I receive the report?
Founding Beta accepts at most five new paid reports per week. We confirm the expected turnaround after payment and source review; every delivery is manually reviewed before it is emailed.
Send the engine a source idea.
Paste a DOI, arXiv ID, public link, title, or short description — or attach a PDF for private storage. Use a paper if you have one; if you are not a researcher, choose a paper or framework you suspect contains a useful mechanism for your problem. The engine generates the report, then a human reviewer inspects it before delivery. We currently accept up to 5 new paid reports per week.
- 1. Choose an offer and send the source.
- 2. Aim any included directed projections; unused slots become scouted cards.
- 3. The engine generates insight cards, fit verdicts, and a Trajectory Tree.
- 4. We review for clarity, citation honesty, weak analogies, and overclaims.
- 5. You receive the reviewed report by email.