TransformText GitHub Action

Transform text with regex replace, base64 encode/decode, URL encode/decode, case conversion, SHA-256 hashing, and more. Zero dependencies beyond @actions/core.

Inputs

Input	Required	Default	Description
text	Yes	—	Input text to transform
operation	Yes	—	Transformation to apply (see operations below)
find	No	—	Regex pattern (for regex_replace)
replace	No	—	Replacement string with $1, $2 groups (for regex_replace)
flags	No	g	Regex flags (for regex_replace)
start	No	0	Start index (for substring)
end	No	—	End index (for substring)
delimiter	No	,	Delimiter (for split)
index	No	—	Element index (for split)

Operations

Operation	Description
regex_replace	Find and replace using regex with capture groups
base64_encode	Encode text to base64
base64_decode	Decode base64 to text
url_encode	URL encode (percent-encoding)
url_decode	URL decode
uppercase	Convert to UPPERCASE
lowercase	Convert to lowercase
trim	Remove leading/trailing whitespace
substring	Extract substring by index
split	Split by delimiter, return all or specific index
reverse	Reverse the string
length	Get string length
lines	Count number of lines
hash_sha256	SHA-256 hash of the text

Output

Output	Description
result	The transformed text

Examples

Regex replace

- uses: skgandikota/TransformText@v1
  id: transform
  with:
    text: "Hello World 2026"
    operation: "regex_replace"
    find: "(\\w+) (\\w+)"
    replace: "$2 $1"

# result: "World Hello 2026"

Base64 encode an API key for headers

- uses: skgandikota/TransformText@v1
  id: encode
  with:
    text: "${{ secrets.API_USER }}:${{ secrets.API_PASS }}"
    operation: "base64_encode"

- uses: skgandikota/FetchUrl@v2
  with:
    url: "https://api.example.com/data"
    headers: '{"Authorization": "Basic ${{ steps.encode.outputs.result }}"}'

Extract domain from URL

- uses: skgandikota/TransformText@v1
  id: domain
  with:
    text: "https://api.example.com/v2/users"
    operation: "regex_replace"
    find: "https?://([^/]+).*"
    replace: "$1"

# result: "api.example.com"

License

MIT