Knowledge should be within reach.
Notanda finds openly available research, records where every item came from, and keeps that record next to your files — so that months later you can still show how you got there.
The tools you search with will keep changing. The record of what you found doesn't have to. Models change. Evidence stays.
Finding the evidence is only half the job.
Most research tools are built to produce an answer, not a record. Which databases were asked, what they actually returned, which version of a full text you ended up with, whether anything has changed since — in most systems that context is never kept as something you can go back to. It exists during the search and not afterwards.
Months later, when you need to show how you got there, you are reconstructing it from memory and a folder of PDFs.
Research you cannot reconstruct is hard to defend.
What Notanda does today
- Searches three sources with different jobs.
- OpenAlex finds candidates and metadata. Europe PMC cross-checks records and supplies full texts. Unpaywall is asked whether an open version exists somewhere else.
- Downloads only what is lawfully open.
- Open access, free licences, legitimate repositories — nothing else.
- Writes down every step.
- Each request, each answer, each failure is appended to a ledger. Provider responses are retained separately, so conflicting source information is not lost when a preferred view is produced.
- Fingerprints every file.
- A SHA-256 checksum for each download, so you can tell later whether the file you have is the file you got.
- Leaves everything on your machine.
- Plain files in a documented structure, not a service you have to keep paying to keep your own research.
We show what exists, and we deliver what is open.
How it works
You say what you are looking for. A question, a topic, a set of criteria.
Notanda asks the sources and keeps the answers — including the ones that contradict each other.
It retrieves what is openly available and takes a fingerprint of every file.
It appends the whole run to the ledger. The search can be read back later: what was asked, what came back, what was reachable, and when.
No step is rewritten. If a provider answers differently next year, both answers are still there.
Where it stands
Status as of 2026-09-12, against release 1.2.0.
Available now
The workflow described above is available in the current beta. Two things are being built: proving that the full history holds up in daily use, and exporting a corpus so it can be moved or restored elsewhere.
Further out
Deliberately not promised yet: watching a corpus for changes over time, taking in results produced by other research tools, and letting AI agents read from the corpus directly.
What Notanda doesn't do
- It doesn't judge quality or truth.
- Notanda records where something came from, not whether it is any good. That judgement stays with you.
- It doesn't get past paywalls.
- No institutional logins, no workarounds. Closed work is recorded as existing, not delivered.
- It doesn't claim to be complete.
- What the sources don't report, Notanda can't show — and a gap is recorded as a gap.
Try it
Notanda works today and is early. Access is handed out one at a time, so that feedback actually gets answered.
Tell us briefly what you are researching and how you work now.