NitHigh confidence

Typo in success toast

"succesfully" should be "successfully".

Screenshot with the text highlighted
Example finding Dashboard GitHub check
MinorMedium confidence

"Sign in" and "Log in" are both used

The same action is labelled "Sign in" on the home page and "Log in" on pricing.

Screenshot and the other places the term appears
Example finding Dashboard GitHub check
01 / 03

Spelling & grammar

Checks every word users can see, including the text developers forget.

Type
Deterministic + AI
Status
[TBD: Live / Beta / Coming soon]
Default
[TBD: On / Off]
Est. cost
[TBD: X credits per run]

What it catches

  • Typos in buttons, headings, toasts, modals and validation messages
  • Grammar slips in new copy
  • Typos in new string literals in the diff

Example finding

A toast message with a typo underlined in red, and the evidence Hydrofoil attaches.

Hydrofoil · Spelling
Suggestion

succesfullysuccessfully

In the diffsrc/components/Toast.tsx:42
  1. 1
    Highlighted textThe misspelt word, marked where users see it.
  2. 2
    Suggested fixThe corrected spelling.
  3. 3
    Strings in the diffReported as file:line instead of a screenshot.

How it works

A dictionary pass finds candidates in all visible text across visited states, then AI filters out product names and jargon.

  1. 01

    Collect visible text

    Across every state the run visited.

  2. 02

    Dictionary pass

    Find candidate typos.

  3. 03

    AI filter

    Drop product names and jargon.

Output

Findings appear in the Hydrofoil dashboard and in your GitHub check, with a severity (blocker, major, minor or nit) and a confidence level. Low-confidence findings are marked Needs reviewand don't fail the check.

Severity
BlockerMajorMinorNit
Confidence
HighMediumLow · Needs reviewDoesn't fail the check
Where it appears
Hydrofoil dashboard GitHub check

Evidence

What each finding includes.

  • Screenshot with the text highlighted, or file:line for strings in the diff.

Known limits

  • Only text in visited states.
  • [TBD: languages supported] [TBD: custom dictionary option]

Configuration

Turn this review on or off for each repository in repository settings.[TBD: add any check-specific options.]

02 / 03

Wording & UX writing

Reviews clarity, tone and consistency of the words in your interface.

Type
AI
Status
[TBD: Live / Beta / Coming soon]
Default
[TBD: On / Off]
Est. cost
[TBD: X credits per run]

What it catches

  • "Sign in" on one page and "Log in" on another
  • Inconsistent capitalisation
  • Vague button labels and unhelpful error messages
  • Leftover lorem ipsum, TODO text, and rendered undefined, NaN or [object Object]

Example finding

Two pages that use different words for the same action, linked together as one finding.

Hydrofoil · Terminology
Same action
  1. 1
    First usageThe term on one page.
  2. 2
    Conflicting usageA different term for the same action elsewhere.

How it works

The agent collects visible text across visited states and reviews it for clarity, tone and terminology.

  1. 01

    Collect visible text

    Across every visited state.

  2. 02

    Review the wording

    For clarity, tone and terminology.

  3. 03

    Link the evidence

    Show where else a term appears.

Output

Findings appear in the Hydrofoil dashboard and in your GitHub check, with a severity (blocker, major, minor or nit) and a confidence level. Low-confidence findings are marked Needs reviewand don't fail the check.

Severity
BlockerMajorMinorNit
Confidence
HighMediumLow · Needs reviewDoesn't fail the check
Where it appears
Hydrofoil dashboard GitHub check

Evidence

What each finding includes.

  • Screenshot with the text highlighted, plus the other places the same term appears.

Known limits

  • Style is partly subjective, so suggestions are usually low severity.
  • Brand voice is only known if you add it to your repo instructions.

Configuration

Turn this review on or off for each repository in repository settings.[TBD: add any check-specific options.]

03 / 03

i18n readiness

Finds what would break when you add or change languages.

Type
Deterministic + AI
Status
[TBD: Live / Beta / Coming soon]
Default
Opt-in
Est. cost
[TBD: X credits per run]

What it catches

  • Hard-coded strings in repos that use an i18n library
  • Missing translation keys
  • Layouts that break when text gets longer
  • Right-to-left layout problems

Example finding

The same button in normal text and in expanded pseudo-text.

Hydrofoil · Pseudo-localisation
Normal text

Fits the button

Pseudo-localised

Text spills outside the button

  1. 1
    Normal textThe layout as designed.
  2. 2
    Pseudo-localised textLonger, accented text that simulates another language.

How it works

Hydrofoil scans the diff for hard-coded strings and missing keys, then re-renders pages with pseudo-localised text to test expansion and right-to-left layout.

  1. 01

    Scan the diff

    Find hard-coded strings and missing keys.

  2. 02

    Pseudo-localise

    Re-render pages with longer text.

  3. 03

    Test layout

    Check expansion and right-to-left.

Output

Findings appear in the Hydrofoil dashboard and in your GitHub check, with a severity (blocker, major, minor or nit) and a confidence level. Low-confidence findings are marked Needs reviewand don't fail the check.

Severity
BlockerMajorMinorNit
Confidence
HighMediumLow · Needs reviewDoesn't fail the check
Where it appears
Hydrofoil dashboard GitHub check

Evidence

What each finding includes.

  • file:line for strings; screenshot of the pseudo-localised page for layout problems.

Known limits

  • Only for repos that use an i18n library.
  • Doesn't judge translation quality. [TBD: confirm]

Configuration

Turn this review on or off for each repository in repository settings.[TBD: add any check-specific options.]

Related reviews

See it on your next pull request

Hydrofoil opens each preview in a real browser and posts what it finds to GitHub.