pre.dev Labs

Private production code.Ready to train on.

For AI labs: hard, verified RL tasks built from its real commits, or the codebases themselves with full git history. Licensed directly from the teams that built them.

Scale

Rounded down, as of 27 Sep 2026

private codebases
10,000+
tokens of private code
50B+
commits of history
2.5M+
new RL tasks every month
1,000+
Two offers

Tasks, or the code itself. Same private corpus.

Both come from private production code licensed directly from the teams that built it. Samples are walked through on a 30-minute call with the founders, after a short form.

RL Tasks

Hard, verified coding tasks

Built from private production code no model has seen, for post-training. One real commit becomes one task.

Per task
A runnable environment at the before-state, the instruction, a graded verifier with continuous reward from 0 to 1, a hidden reference patch, QC evidence, and solver trajectories.
Checks
The reference solution scores at least 0.95 in a cold container; doing nothing and cheat probes score at most 0.1; fairness, mutation and consistency checks pass. Only tasks that solver runs score from 0.1 to 0.8 ship.
Exposure
Every source repo passes an exposure check. Every bundle ships a file-tree manifest (sha256 per file plus a root hash) so you can check overlap with your own corpus.
Pricing
Priced per engagement, quoted after a scoping call.
Delivery
Harbor task format, as a tar.gz with its sha256. 1,000+ new tasks every month.
Licensing
Non-exclusive by default.

Codebases

The code itself, with its history

Licensed private codebases with full git history, cleaned and delivered for training.

Per repo
  • Full git history, all branches and tags, rewritten to remove secrets and personal data, with authors pseudonymized
  • Pull requests with diffs and reviews, as JSON and in GitHub API shape
  • A README explaining how to rebuild each pull request as a task
  • Build checks
Cleaning
  • Three secret scanners over every file version
  • Emails, phone numbers and profile URLs removed
  • Authors mapped to pseudonyms, the origin organization name rewritten
  • Archives, binaries, vendored dependencies, IDE files, build output and bulk data files stripped
Exposure
Every repo passes an exposure check before listing: not public, not a fork, not archived, not a published package.
Pricing
Priced per token: one rate for source code, one for git history. Quoted on the call.
Delivery
Delivered through expiring signed links.
Licensing
Licensed directly from the teams that built the code. Non-exclusive by default.

Book an intro call

30 minutes with Arjun and Adam, the founders. Tell us what you're training, and we'll scope RL Tasks, codebases, or both.

Tell us what you're training

Then pick a time. We walk through samples on the call and scope a set with your team.

Interested in (pick any)

Goes straight to the founders.