Your business has been generating training data for years. Find out what it’s worth.
CorpusReady takes your records from “somewhere in a spreadsheet” to a de-identified, documented data pack with an indicative value score, ready to submit to an AI data marketplace.
A corpus is what AI companies call the data they learn from. CorpusReady gets yours ready to license.
- Operated by Trove Distribution Limited, UK
- Raw uploads deleted 24 hours after the pack is written
- No LLM ever receives raw data
- Flags issues, not legal advice
Prepares packs for
- Troveo
- Opendatabay
- FileYield
See what you hold before you share anything
Every screen below is the product as it runs today, with demo business data.
Inventory
Upload a CSV or XLSX export, or answer 14 questions about what you hold. CorpusReady checks for data it will not process before anything else happens.
Rights check
A rights matrix marking each source green, amber or red, with the clause that would need updating. A flag, not legal advice.
De-identification
Deterministic detection and redaction of names, emails, phone numbers, addresses, IDs and card numbers, plus a k-anonymity check on quasi-identifiers. You review a before-and-after sample.
Five steps, one artefact each
Stop after step 1 for free. Steps 2 to 5 are the paid pack.
Inventory
FreeUpload a CSV or XLSX export, or answer 14 questions about what you hold. CorpusReady checks for data it will not process before anything else happens.
Output: A data inventory: sources, record counts, date ranges, fields and likely category. Free.
Rights check
Data PackA fixed questionnaire per source: who created the records, what your customer terms and privacy notice said, whether any contract or licence restricts sharing.
Output: A rights matrix marking each source green, amber or red, with the clause that would need updating. A flag, not legal advice.
De-identify
Data PackDeterministic detection and redaction of names, emails, phone numbers, addresses, IDs and card numbers, plus a k-anonymity check on quasi-identifiers. You review a before-and-after sample.
Output: The cleaned dataset and a redaction log.
Provenance pack
Data PackDocumentation built from metadata and the redacted sample only: origin, collection method, date coverage, field dictionary, de-identification method, licence basis and your attestation.
Output: A provenance PDF and a machine-readable JSON manifest.
Score and route
Data PackA heuristic value score with an indicative price band, then submission to a matching AI data marketplace with the pack attached. This is the only step where anything leaves CorpusReady, and only when you choose to send it.
Output: A scorecard and a submission receipt.
Three ways to get data in front of AI buyers
What changes depending on who does the work.
| Question | CorpusReady | A data broker | Doing it yourself |
|---|---|---|---|
| Who holds the data | You do. Raw uploads are deleted 24 hours after the pack is written, 30 days after payment if a paid pack is not run, or 7 days after upload if no pack is bought. | The broker, once you hand it over. | You do. |
| Cost | Free inventory. Data Pack $149 per dataset, or Pro $39 per month. | Set by each broker. | Your own time, plus any legal or technical help you pay for. |
| Time to a pack | The inventory runs as soon as your file is checked. The pack follows the four paid steps, at your pace. | Depends on the broker. | Depends on how much time you can give it. |
| Legal exposure | You decide what to license. The rights check flags issues per source. This is not legal advice. Consult a lawyer before licensing. | Set by your contract with the broker. | You check rights and de-identification yourself. |
| What you keep | Your data, and the de-identified pack, redaction log, manifest and scorecard. CorpusReady does not buy or resell data. | Whatever the contract leaves you. | Everything, including all of the work. |
What we won’t process
CorpusReady refuses these at upload. It detects them, stops, explains why and deletes the file.
| Category | Such as |
|---|---|
| Health data | Diagnoses, prescriptions, medication or patient records |
| Biometric data | |
| Data about children | |
| Financial account data | Card numbers and IBANs |
| Government ID numbers | Malaysian NRIC, passport, UK National Insurance and US Social Security numbers |
Pricing
Start free. Pay only when you want a pack for a dataset.
Free inventory
$0
Step 1. Upload a CSV or XLSX file, or answer 14 questions. See what you hold: sources, record counts, date ranges, fields and likely personal data.
Start your free inventoryData Pack
$149per dataset
Steps 2 to 5 for one dataset: rights check, de-identification, provenance pack, value score and submission to a matching marketplace.
Start free, upgrade laterPro
$39per month
Unlimited packs and re-runs. For businesses with several datasets or data that grows every month.
Start free, upgrade laterWhat each plan includes
| Step | Free inventory | Data Pack | Pro |
|---|---|---|---|
| 1. Inventory | Included | Included | Included |
| 2. Rights check | Not included | One dataset | Unlimited |
| 3. De-identification | Not included | One dataset | Unlimited |
| 4. Provenance pack | Not included | One dataset | Unlimited |
| 5. Value score and submission | Not included | One dataset | Unlimited |
| Price | $0 | $149 per dataset | $39 per month |
All prices are charged in US dollars. Malaysian visitors see an approximate ringgit figure for guidance: Data Pack approx. RM 490, Pro approx. RM 129 per month. Any value score is indicative, not an offer. CorpusReady does not buy or resell data and cannot guarantee a sale.
Why I built CorpusReady
I co-founded Hedon, a premium motorcycle helmet brand, and ran it as CEO for thirteen years. We started pure B2B through distributors, moved to selling direct, and ended up shipping to riders in more than 70 countries with a team of about thirteen people.
A business like that generates data every day without meaning to: thirteen years of orders, customers, sizing, returns, dealer sell-through, product feedback across markets. When AI companies started paying for training data, I looked at what we held and realised there was no sensible way for a company our size to find out what it was worth, or to get it into a shape a buyer could legally use. The brokers wanted a Fortune 500 contract. The marketplaces wanted a finished, documented, de-identified dataset. Nobody helped you get from “it’s in a spreadsheet somewhere” to that.
CorpusReady is the tool I wished we’d had. It inventories what you hold, flags the rights questions you need to ask, removes the personal data, documents where everything came from, and gives you an honest indicative figure before you decide whether to approach a marketplace. You keep the data. We never see it raw. Nothing goes anywhere unless you send it.
— Reg Flint, founder. CorpusReady is operated by Trove Distribution Limited, UK.
Questions
Does CorpusReady buy my data?
No. CorpusReady prepares your data and introduces it to existing AI data marketplaces, such as Troveo, Opendatabay and FileYield. It never buys, resells or licenses your data, and it cannot guarantee a sale.
What is the value score?
A heuristic based on volume, time span, free text, language, domain and structure. The price band is indicative, not an offer. Marketplaces set their own terms.
What happens to my raw file?
Raw uploads are deleted 24 hours after the pack is written, 30 days after payment if a paid pack is not run, or 7 days after upload if no pack is bought. Files we refuse are deleted straight away. A scheduled job runs every hour and logs each deletion.
Does any AI model see my customers' data?
No language model ever receives raw customer data. Detection and redaction are deterministic. A model may draft parts of the provenance document from metadata and redacted samples only.
Is the rights check legal advice?
No. It flags issues so you know what to ask. This is not legal advice. Consult a lawyer before licensing.
Who is CorpusReady for?
Small businesses in the UK and Malaysia with years of records: orders, support tickets, listings, jobs, CRM notes or invoices.
What files can I upload?
CSV and XLSX files up to 200 MB. If your data is somewhere else, you can answer the questionnaire instead and get an inventory without uploading anything.
Start with a free inventory
What you hold, how many records, which years, and which fields contain personal data.