AI data vendor

pre.dev: what they buy and what they pay

pre.dev is profiled here as an AI data vendor[1]. Data types cited in this profile include: Code, Git history, Pull requests/code reviews[2].

Last verified November 3, 2026

Quick facts

Quick facts about pre.dev
TypeAI data vendor[1]
StatusActive[2]
Websitehttps://pre.dev[3]
HQMailing address per Terms: 2140 S Dupont Hwy, Camden, DE 19934, United States[4]
FoundedNot published
OwnershipLegal entity: PREDEV CORP[4]Founders: Arjun and Adam[2]
ProgramCode licensing ("License your codebases")[5]https://pre.dev/code-licensing[5]
PriceNot published
Pricing basisLab-facing codebase pricing: per token, one rate for source code and one for git history, quoted on the call[2]
Payment timingNot published
How to startCreate an individual or organization account at pre.dev/code-licensing[5]

What they buy

What they buy: pre.dev
Data typesCode, Git history, Pull requests/code reviews[2]
ExcludedPublic repos, forks, archived repos and published packages are excluded[2]Archives, binaries, vendored dependencies, IDE files, build output and bulk data files stripped[2]
Intended useTraining (codebases)[2]RL post-training tasks[2]

Price and payment

Price and payment: pre.dev
PriceNot published
Pricing basisLab-facing codebase pricing: per token, one rate for source code and one for git history, quoted on the call[2]
Payment timingNot published
Seller feesNot published
Platform's cutNot published
Referral bountyNot published

Deal terms

Deal terms: pre.dev
Deal structureLicense (lab-facing licenses non-exclusive by default)[2]
ExclusivityNon-exclusive by default (lab-facing)[1]
TermNot published
Field of useNot published
Onward transferpre.dev turns licensed code into RL tasks and also licenses the codebases themselves to AI labs[2]
Seller keeps ownershipNot published

Privacy and data handling

Privacy and data handling: pre.dev
De-identificationpre.dev runs three secret scanners, removes emails/phone numbers/profile URLs, pseudonymizes authors and rewrites the origin organization name[2]
Raw data deletionNot published
CertificationsNot published
Personal data handlingGit history rewritten to remove secrets and personal data; authors pseudonymized[2]Delivery to buyers through expiring signed links[2]

Process

Process: pre.dev
StepsNot published
TimelineNot published
Sample before offerNot published
How to startCreate an individual or organization account at pre.dev/code-licensing[5]

Who they sell to

Who they sell to: pre.dev
Sells toFrontier AI labs (unnamed)[6]

Watch for

None found in public reporting as of November 3, 2026.

Sources

  1. RL Tasks: Verified Coding Tasks for Post-Training | pre.dev, pre.dev. Accessed October 30, 2026. (company's own site)
  2. pre.dev Labs: RL Tasks and Codebases for AI Labs, pre.dev. Accessed October 30 and November 3, 2026. (company's own site)
  3. pre.dev | Long Horizon Coding Agent, pre.dev. Accessed October 30, 2026. (company's own site)
  4. Terms and Conditions | pre.dev, pre.dev. Published November 3, 2026. Accessed November 3, 2026. (company's own site)
  5. License your codebases | pre.dev, pre.dev. Accessed October 30, 2026. (company's own site)
  6. pre.dev llms.txt, pre.dev. Accessed October 30, 2026. (company's own site)

Last verified November 3, 2026