Updated November 2026
Grepped vs Collinear: which should see your company's code and records?
If your company is not a software shop, Grepped's form lists the wider range of industries, from accounting and construction to logistics and legal, and it states no headcount minimum. Collinear says more about what it wants, mainly private repos with pull requests, tickets and CI logs, plus sectors such as healthcare operations, banking, legal and construction, and it prefers teams of about 50 people or more, though smaller teams with deep records can work.1,2,3,4
Side by side
Grepped and Collinear, row by row
| Topic | Grepped | Collinear |
|---|---|---|
| What it takes | Messages, codebases, documents, business records and task histories5 | Private repos with pull requests, tickets and CI logs, coding-agent sessions, business app history and database dumps2 |
| Who qualifies | No headcount minimum stated; many industries on its form1 | About 50+ people as best fit, smaller teams with deep records can work; English data, US, Canada, UK and EU, Australia or New Zealand4 |
| What stays out | Whatever you choose; any system, folder or record type6 | Whatever you prefer not to send; separately, it says it does not need AI-generated code, thin layers on public frameworks, loose documents or mostly non-English data7,8 |
| How access works | Read-only access to approved sources6 | A one-time export; any access granted can be removed once the transfer is done9 |
| Names and personal details | Names, emails and phone numbers removed6 | Removed after export; names become consistent pseudonyms7 |
| After the work is done | Originals deleted after processing; retention of the processed copy set in the agreement6 | No retention period or deletion promise on its page10 |
| What it does with the data | Reviews, packages and resells it to train models; Grepped says it is not sold to your competitors5,11 | Builds tasks and simulated environments for AI agents, expanded synthetically12 |
| Who you are dealing with | Grepped AI, Inc., Delaware, San Francisco; its site names no people13,14 | Collinear AI, Inc., Sunnyvale; CEO Nazneen Rajani15,16 |
If you write software
What makes a codebase interesting, and what doesn't
Collinear is unusually open about what it values. Fifteen or more people at peak, several years of history, pull requests tied to their tickets, code written by people, and the same work visible across several tools all raise the price. A public framework with a thin layer on top, or code from a very small team, is on its list of what it does not need.17,8
Grepped takes codebases as one part of a wider set of records, from Slack and Gmail to GitHub, Jira, Salesforce, QuickBooks and Zendesk. Collinear also takes business app history from many of the same tools, so the difference is how much it leads with code.5,2
Control
What you can hold back, and what to ask
Both let you hold things back. Grepped says any system, folder or record type can be left out, and Collinear says you can hold back whatever you prefer not to send. Collinear also signs a mutual NDA before samples on request.6,7,18
Grepped's site code calls its own estimator "deliberately optimistic", so don't lean on any figure it shows you. Collinear says ownership stays with you under a license, but its page does not say who performs the export and mentions no retention, deletion, exclusivity or fees, so those are your first questions, and ask Grepped about exclusivity and fees too.19,9,7,10
Before either one sees a repo, check what you are free to share. Grepped itself says some customer and supplier agreements restrict how data can be used, which matters most for client work and agency code.20
FAQ
Questions, answered with sources
Every business is different, so this one is best talked through with the buyer you're working with. Their team goes over it with you before anything is signed.
Before you sign
Questions to talk through with your buyer
Every business is different, so these are best talked through with the buyer you're working with. Their team goes over them with you before anything is signed.
Questions for this kind of business
- Which repos contain client work, and do those client agreements let us license it?
- Who owns this data: the company, an estate, or someone else?
- Do customer, supplier or platform contracts, NDAs or API terms restrict this?
Questions for any buyer
- Is the license exclusive, and for how long?
- Can you sublicense the data, and to whom?
- Which AI labs will receive it, and will you tell us?
- How long do you keep it, and how is it deleted?
- What is your liability cap, and who indemnifies whom?
- Will the contract ban re-identification of people in the data?
- Who removes personal data, by what method, and do we approve it first?