docs(faq): answer why a Pro/Max subscriber gets billed for API tokens - #32
Open
santifer wants to merge 2 commits into
Open
docs(faq): answer why a Pro/Max subscriber gets billed for API tokens#32santifer wants to merge 2 commits into
santifer wants to merge 2 commits into
Conversation
…ertas
search-ops fixed the canon on 2026-08-11 after 18 days without objection: the
report is eight blocks, A through H, with H conditional. This is the group-2
sweep measured on 2026-07-30, executed in one pass.
Verified against the core before writing (modes/oferta.md, commit 350ab739):
eight `## X)` headings, and `## H) Draft Application Answers` gated on
"only if score >= 4.5". The site was carrying three different answers at once
— A-F, A-G and "seven-block" — depending on which file you opened.
17 files. Descriptions and prose in modes/{index,oferta,auto-pipeline} across
EN/ES/FR; the Block bullet list in docs/index across EN/ES/FR, which was
missing H entirely; schema.ts (`Block A-G evaluation` and "seven-section");
llms.txt, llms-full.txt and comparisons.json. translationHash re-stamped on
all eight touched translations.
Also lands the disambiguation search-ops asked for, in the passages engines
quote (methodology and llms-full):
oferta (singular) — ONE listing, five-dimension holistic rubric, no formula
ofertas (plural) — MULTIPLE offers, ten-dimension WEIGHTED matrix
Both are true and both are public, and their mode names differ by one letter.
That collision is the actual root of the "ten weighted parameters" that
reached press coverage, not a bad translation. Naming the mode every time a
dimension count appears is now doctrine; the fix is to state both, not to ban
the word "weighted", which would delete a true fact about a real mode.
Verified on a production build: zero A-F/A-G left on any served surface,
eight Block bullets on /docs.md, Block H present in all three locales.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Triple signal per search-ops: the #1 community question on X this week (answered by hand by @careeropsHQ), a real user burning in Discord who asked it twice with no reply, one of them in Russian, and long-standing recurring confusion. Every fact verified against code by career-ops-maintainer before writing, and one answer changed the register of the doc: the precedence is NOT ours. Claude Code decides, ANTHROPIC_API_KEY wins, and career-ops only detects it and warns. So this is written as "your CLI does X, and here is how you check", not "career-ops does X". Source: doctor.mjs:71-110 checkBillingSource(), frozen by tests/doctor-billing-source.test.mjs. The command goes first, above the explanation. The maintainer's point is that `node doctor.mjs` already answers this today and nobody knows: that is not missing documentation, it is correct documentation nobody links. The doctor's warning is quoted verbatim rather than paraphrased. Deliberately NOT written: that exporting an empty ANTHROPIC_API_KEY sidesteps the problem. It is true of our doctor (an empty string is falsy at doctor.mjs:72) but unverified on the CLI side, and a cost doc is the worst place to guess: if someone believes us and gets charged, that is on the doc. Also deliberately declared rather than filled: career-ops does not check Codex, OpenCode, Copilot, Gemini or Qwen, so the page says each CLI sets its own rule and points at their documentation. A declared gap beats a guessed one. FAQ entry in EN/ES/FR plus the faq-data.ts mirror that backs the FAQPage schema, and a warning callout in getting-started right before the first evaluation, which is the moment the money starts. translationHash re-stamped on all four translations. Stacks on #31 (both touch content/docs/index across the three locales). Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
santifer
force-pushed
the
fix/rubric-canon-a-h
branch
from
September 10, 2026 20:48
727af1b to
ad677bc
Compare
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
1 participant
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.
Why now
Triple signal, per search-ops:
The fact that changed the register of the doc
Every claim was verified against code by career-ops-maintainer before writing. One answer reframed the whole page:
So this is written as "your CLI does X, and here is how you check", not "career-ops does X". Source:
doctor.mjs:71-110checkBillingSource(), frozen bytests/doctor-billing-source.test.mjs.The command goes first
The maintainer's observation, which is the real insight:
So
node doctor.mjsis the first line of both surfaces, above the explanation. And the doctor's warning is quoted verbatim rather than paraphrased.Two things deliberately NOT written
That exporting an empty
ANTHROPIC_API_KEYsidesteps it. True of our doctor (an empty string is falsy atdoctor.mjs:72, the maintainer verified it) but unverified on the CLI side. A cost doc is the worst place to guess: if someone believes us and gets charged, that is on the doc.Any rule for the other CLIs. career-ops does not check Codex, OpenCode, Copilot, Gemini or Qwen. The page says each CLI sets its own rule and points at their docs. A declared gap beats a guessed one.
Changes
content/docs/faq.{mdx,es,fr}src/lib/faq-data.tsFAQPageschemacontent/docs/index.{mdx,es,fr}translationHashre-stamped on all four translations.Verified on a production build
Why is career-ops billing my API creditspresent in the renderedFAQPageJSON-LD.🤖 Generated with Claude Code