FoodProt AI

From research question to verified dataset

Search, screen, extract, verify and export research evidence. Every value keeps the page, table or figure it came from, and a person approves it before it counts.

Create account Sign in How it works

One verified valueApproved

Property: Least gelation concentration
Reported: 12.0 % w/v
Standardised: 12.0 % w/v
Conditions: pH 7.0 · 0.1 M NaCl · 90 °C, 30 min
Source: Table 3, page 4
Reviewed by: A named reviewer, with the date

An illustration of the record layout, not data from a paper.

How it works

Six stages, from question to dataset

Each stage keeps a record of what was done and by whom. Nothing enters the dataset until a person approves it.

1
Discover
Search the literature indexes or import your own references. Duplicates across sources are merged.
Automatic
2
Screen
Each paper is checked against your criteria, with a suggested decision and its reason. You decide.
Suggested, you decide
3
Full text
Open-access full text is collected automatically. For the rest, use your library or upload the file.
Automatic
4
Extract
Values are taken from text, tables and figures, each with the passage it came from and its original unit.
Suggested, you decide
5
Review
Each value is shown beside its source. A reviewer accepts, corrects or rejects it, and that is recorded.
Decided by a person
6
Dataset
Approved rows export as CSV, XLSX or JSON with reported and standardised values, conditions and sources.
Automatic

Why the numbers hold up

Evidence first, and a person in charge

Traced to the source

Each value links to the article, page and table, figure or sentence it was taken from.

Original and standardised

The value and unit as printed are kept beside the standardised value. Ambiguous units are not converted.

Reviewed by a person

A reviewer accepts, corrects or rejects each value. Their decision is recorded and never overwritten.

Honest about limits

When the evidence is unclear the value is flagged for review instead of being guessed.

Where the literature comes from

Search what can be searched. Bring the rest yourself.

Search

PubMed, OpenAlex, Crossref, Semantic Scholar and Europe PMC are public metadata APIs, documented for programmatic use. This platform queries all five directly, at the rate each one publishes. Every search is saved with its query, filters and date.

Import

Google Scholar, ResearchGate and MDPI (a publisher, not a database) are never queried and never scraped. Search them yourself, import what you export, or follow a link this platform builds for you to open. Paste DOIs or PubMed IDs, or import RIS, BibTeX, CSV or XLSX from any database you have access to.

Full text

Open-access copies are collected automatically. Anything else comes through your own library access or a file you upload.

Scopus and Web of Science are optional official APIs, off by default. A platform owner whose institution holds the entitlement switches one on and enters the credential; until then nothing is sent to it. Paywalls and subscriptions are never bypassed. How full text is collected

Help

Read how it works before you sign in

All guides

Start with a private project

Create an account, set your question and criteria, and invite colleagues when you are ready.

Create account Sign in

FoodProt AI

From research question to verified dataset

Search, screen, extract, verify and export research evidence. Every value keeps the page, table or figure it came from, and a person approves it before it counts.

Create account Sign in How it works

One verified valueApproved

Property: Least gelation concentration
Reported: 12.0 % w/v
Standardised: 12.0 % w/v
Conditions: pH 7.0 · 0.1 M NaCl · 90 °C, 30 min
Source: Table 3, page 4
Reviewed by: A named reviewer, with the date

An illustration of the record layout, not data from a paper.

How it works

Six stages, from question to dataset

Each stage keeps a record of what was done and by whom. Nothing enters the dataset until a person approves it.

1
Discover
Search the literature indexes or import your own references. Duplicates across sources are merged.
Automatic
2
Screen
Each paper is checked against your criteria, with a suggested decision and its reason. You decide.
Suggested, you decide
3
Full text
Open-access full text is collected automatically. For the rest, use your library or upload the file.
Automatic
4
Extract
Values are taken from text, tables and figures, each with the passage it came from and its original unit.
Suggested, you decide
5
Review
Each value is shown beside its source. A reviewer accepts, corrects or rejects it, and that is recorded.
Decided by a person
6
Dataset
Approved rows export as CSV, XLSX or JSON with reported and standardised values, conditions and sources.
Automatic

Why the numbers hold up

Evidence first, and a person in charge

Traced to the source

Each value links to the article, page and table, figure or sentence it was taken from.

Original and standardised

The value and unit as printed are kept beside the standardised value. Ambiguous units are not converted.

Reviewed by a person

A reviewer accepts, corrects or rejects each value. Their decision is recorded and never overwritten.

Honest about limits

When the evidence is unclear the value is flagged for review instead of being guessed.

Where the literature comes from

Search what can be searched. Bring the rest yourself.

Search

Import

Full text

Open-access copies are collected automatically. Anything else comes through your own library access or a file you upload.

Help

Read how it works before you sign in

All guides

Start with a private project

Create an account, set your question and criteria, and invite colleagues when you are ready.

Create account Sign in

From research question to verified dataset

Six stages, from question to dataset

Discover

Screen

Full text

Extract

Review

Dataset

Evidence first, and a person in charge

Traced to the source

Original and standardised

Reviewed by a person

Honest about limits

Search what can be searched. Bring the rest yourself.

Search

Import

Full text

Read how it works before you sign in

Getting started in FoodProt AI

Reviewing extracted values

Units and standardisation

Exports and the audit trail

Start with a private project

From research question to verified dataset

Six stages, from question to dataset

Discover

Screen

Full text

Extract

Review

Dataset

Evidence first, and a person in charge

Traced to the source

Original and standardised

Reviewed by a person

Honest about limits

Search what can be searched. Bring the rest yourself.

Search

Import

Full text

Read how it works before you sign in

Getting started in FoodProt AI

Reviewing extracted values

Units and standardisation

Exports and the audit trail

Start with a private project