See what the research actually makes possible.
Connect evidence across technical literature to identify emerging capabilities, commercial applications, and the assumptions that still need to be proven.
Three ratings, named assumptions, the main risk.
Evidence strength, technical readiness and commercial access are rated separately, each with its reason; there is no overall score to hide behind. The assumptions and the main risk are part of the card, not a footnote.
Measurement-based methane inventories as a service for gas exporters
A recurring service that flies aerial surveys over an operator's assets, adds modelled estimates for sources below detection, and delivers a site-level inventory with uncertainty bounds and sample-size guidance, packaged for voluntary measurement standards and possible EU import equivalence. C-008-0001C-008-0004C-015-0003C-033-0004+3
- Who buys
- Likely buyers: producers and LNG exporters selling into Europe, and companies seeking top-level measured reporting; assumed annual contracts.
- Why now
- UNEP reports one-third of production reports or will soon report with real-world measurements; the EU regulation introduced an import MRV equivalence framework; measured regional inventories exceeded official ones.
Why these ratings
- Evidence: High
- Several regional inventories and blind-tested aircraft support the method and the undercount; demand itself is not measured.
- Readiness: High
- The hybrid measured-plus-modelled protocol has been run for whole provinces with commercial aircraft.
- Access: Medium
- Our judgement: demand hinges on EU rules whose timing may slip and on voluntary standards.
Assumptions
- EU equivalence rules end up accepting aerial measurement-based inventories
- Operators will pay for a number that may raise their reported emissions
- Aircraft survey cost per site stays low enough for annual repeat
Main risk. A delayed or softened EU import standard removes the compliance pull, leaving only voluntary buyers.
It rates things down
A low rating stays low and says why. This one, from the finance investigation, is rated Low on commercial access.
Stress-testing service for fraud and AML model validation
- What to build
- A validation kit that runs a bank's model against synthetic scenarios: rarer crime, new laundering mixes, time drift, scarce or noisy labels, and group fairness. It outputs a model-risk report.
- Who buys
- Model risk management and internal audit teams at banks, and vendors needing evidence for client procurement.
- Why now
- Open generators now simulate label noise, drift and partial visibility, and they show models collapsing when the laundering mix changes. Reviews note no shared benchmarks with operational metrics.
Main risk. Without proof that synthetic results predict real performance, buyers may treat the report as a checkbox.
Compress the path from technical evidence to a credible commercial hypothesis.
The problem is not generating possible applications. The problem is determining whether the technical capability is actually supported, what prerequisites remain, which applications fit those constraints, and whether the opportunity deserves more diligence.
An investigation answers those in order, and keeps the evidence for each answer one click away.
The test of a good investigation here is that you find yourself saying: This saves me the first several days of an investigation.
Contradictions stay on the page
Where the sources disagree, the investigation says so and says why the results may not be comparable. It does not pick a side it cannot support.
A vendor-reported method document gives a detection limit of 0.5 kg/hr at 90% probability of detection at the test centre. A second vendor reports 96.7% correct localisation there. On operating sites, point-sensor networks reported zero during 38-86% of blind releases, and their rates did not track the true ones. The difference is test conditions against field conditions. It is unresolved.
The discipline behind the numbers
An investigation looks far wider than it reads. Every source it considered is listed with its status and the reason, so the shortlist can be challenged too.
63 shortlisted sources could not be fetched (no open copy, or unreadable) and are listed as such. Every source read in full is cited in the report, with why it was added.
See every source it looked atHow it works
- 01
Supply the research
Upload 25–100 papers and tell Between Papers what you're trying to understand.
- 02
Build the evidence
Technical findings, experimental conditions, limitations, agreements, and contradictions are extracted and connected across the corpus. Each claim is checked against the passage it came from.
- 03
Investigate what they enable
Findings are connected into emerging technical capabilities, including potentially useful connections from adjacent fields and, where it helps, sources from outside your corpus.
- 04
Explore commercial implications
Capabilities become potential applications and business opportunities, while evidence, assumptions, and uncertainty remain visible.
Questions
Can I verify the conclusions?
Yes. Important conclusions can be traced back through supporting findings to source passages.
Does Between Papers treat every paper as equally credible?
No. Conflicting evidence, different experimental conditions, limitations, and uncertainty remain visible.
Does a commercial application mean the technology will definitely work?
No. Between Papers identifies evidence-backed possibilities and explicitly surfaces what remains unproven.
Not your job? Find your next product in the research. · What could your research become?
Compress the first days of a technical-commercial diligence.
Read a real investigation first, then run one on the technology you are evaluating.
The flagship: cheaper batteries for homes and the grid, 25 seed papers, 396 sources looked at.