Name camoufox only, and stop describing video as half-shipped - #11
Merged
Conversation
Two unrelated corrections to public-facing copy. ── Holo does not belong in a public repo The end-to-end browser benchmark published a second row — "Holo 152.0.3, 1/3" — in README.md and docs/performance.md. Holo is the private camoufox fork; naming it here, with a version number and a measured solve rate, discloses both that it exists and how it performs. The public integration is camoufox, and that is the only browser this repo should name. THE ROW IS REMOVED, NOT RELABELLED. Those numbers were measured on Holo, and moving them under the camoufox heading would be publishing a benchmark figure for a run that never happened. The surrounding prose leaned on there being two rows — "both runs", "the difference between the two rows" — so it is rewritten for one. The IP-reputation caveat survives on its own merits: it was always the more useful half, and it is true of a single datacenter run too. Checked before editing: the published npm and PyPI descriptions for 2.6.0 carry no mention, and neither does captchakraken.com. It reached the public repo in 8928149 and it is in that commit's history, which a file edit cannot reach — see the PR for what that does and does not expose. ── Video challenges have shipped, in both halves The feature table said "🟡 Video challenges — driver support ships; the adapter `setup.sh` installs skips them", and three other places sold video as a reason to watch the repo. All of that predates v1.2. `setup.sh` installs `CaptchaKraken-Lora-v1.2` / `captcha-v12`, pinned to models.json `latest` by python/tests/test_setup_sh_model_ids.py. That model is prompt generation 2, and generation 2 is the generation that HAS a video prompt — generation 1's is `None`, which is exactly why the old caveat was true when it was written. The hosted API serves the same `captcha-v12`, and docs/hosted-api.md already bills a video response at 10 credits. Both ports get it: types.ts carries `videoSolveEnabled` / `videoBurstDurationMs`, and the JS driver runs the solve through the bundled Python CLI. roadmap.md already said 🟢 shipped and was the only file that had kept up, though it opened with "the driver half has shipped" — accurate when the model half was outstanding, misleading now that it is not. The roadmap pointers are corrected while here: they promised "smaller models", which is not on the roadmap. The one ⚪ planned item is more captcha types.
Owner
Author
✓ Tier 3 driver-gate — passAggregate: 0.807 · 6/10 families passing · 17 pair(s) not solved
|
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Two unrelated corrections to public-facing copy.
Holo does not belong in a public repo
The end-to-end browser benchmark published a second row —
Holo 152.0.3 | 1 / 3— inREADME.mdanddocs/performance.md. Holo is the private camoufox fork; naming it here, with a version number and a measured solve rate, discloses both that it exists and how it performs. The public integration is camoufox, and that is the only browser this repo should name.The row is removed, not relabelled. Those numbers were measured on Holo, and moving them under the camoufox heading would be publishing a benchmark figure for a run that never happened. The prose leaned on there being two rows ("both runs", "the difference between the two rows"), so it is rewritten for one. The IP-reputation caveat survives — it was always the more useful half.
Exposure, checked rather than assumed
captchakraken@2.6.0READMEcaptchakraken@2.6.0description/,/docs,/pricing,/contact)8928149, and this PR cannot reach itThat last row is the one worth a decision. The mention is public in
8928149and stays public unless that history is rewritten — which on a repo that is a submodule of CaptchaKrakenFinetune, and that people have cloned, is a disruptive operation. Flagging it rather than doing it.Video challenges have shipped, in both halves
The feature table said "🟡 Video challenges — driver support ships; the adapter
setup.shinstalls skips them", and three other places sold video as a reason to watch the repo. All of that predates v1.2:setup.shinstallsCaptchaKraken-Lora-v1.2/captcha-v12, pinned tomodels.jsonlatestbypython/tests/test_setup_sh_model_ids.pyNone, which is why the old caveat was true when writtencaptcha-v12, anddocs/hosted-api.mdalready bills a video response at 10 creditstypes.tscarriesvideoSolveEnabled/videoBurstDurationMs, and the JS driver runs the solve through the bundled Python CLIroadmap.mdalready said 🟢 shipped and was the only file that had kept up — though it opened with "the driver half has shipped", accurate when the model half was outstanding and misleading now.Roadmap pointers corrected while here: they promised "smaller models", which is not on the roadmap. The one ⚪ planned item is more captcha types.
All relative links verified to still resolve.