Templates

Templates control how Claude post-processes your transcribed speech. Each template has a name (shown in the Quick Pick menu) and a prompt (the instruction sent to Claude).

Built-in Templates

Verba ships with 8 default templates:

Template	Description	Context-Aware
Freitext	Cleans up transcript: removes filler words, smooths sentences, fixes errors
Commit Message	Converts transcript to Conventional Commits format
JavaDoc	Generates a JavaDoc comment block with @param, @return, @throws
Markdown	Structures transcript with headings, lists, and emphasis
E-Mail	Formats transcript as a professional email
Code Comment	Generates a code comment based on transcript and surrounding code	Yes
Explain Code	Answers questions about code using transcript and codebase context	Yes
Claude Code Prompt	Converts transcript into a prompt for Claude Code with file references	Yes

Template Auto-Reuse

On first dictation, Verba shows the Quick Pick menu for template selection. After that, your last template is automatically reused — press Cmd+Shift+D to start recording immediately.

To switch templates, press Cmd+Alt+T (Mac) / Ctrl+Alt+T (Windows/Linux). The status bar always shows the currently active template.

Context-Aware Templates

Templates with contextAware: true trigger a semantic code search before sending the transcript to Claude. Verba searches your codebase for relevant files, classes, and functions, and includes them as context snippets in the prompt.

This requires a context provider:

• grepai (recommended) — Install grepai and run grepai init in your project.
• OpenAI Embeddings — Run the command Verba: Index Project to build a local index.

Configure the provider in Settings:

{
  "verba.contextSearch.provider": "auto",
  "verba.contextSearch.maxResults": 5
}

The auto setting uses grepai if installed, otherwise falls back to OpenAI Embeddings.

Custom Templates

Define custom templates in settings.json:

{
  "verba.templates": [
    {
      "name": "Free Text",
      "prompt": "Clean up the transcript: remove filler words, smooth broken sentence starts, fix transcription errors. Keep the original language and meaning. Return only the cleaned text."
    },
    {
      "name": "Code Review",
      "prompt": "Convert this transcript into structured code review feedback with bullet points for issues found and suggestions. Keep the original language.",
      "contextAware": true
    }
  ]
}

Template Properties

Property	Type	Required	Description
name	String	Yes	Display name in the Quick Pick menu
prompt	String	Yes	System prompt sent to Claude
contextAware	Boolean	No	If true, includes code context from semantic search

Writing Good Prompts

• Be specific about the desired output format
• Tell Claude to "keep the original language" if you dictate in different languages
• End with "Return only the [result]" to avoid explanatory text in the output

Glossary Interaction

If you have defined a glossary, the glossary terms are automatically included in every template's Claude prompt. This ensures that product names, technical jargon, and abbreviations are preserved exactly — regardless of which template you use. No additional template configuration is needed.

Course Correction

Verba automatically detects and removes self-corrections in your speech. If you say "let's meet tomorrow, no wait, on Friday at ten", only "let's meet on Friday at ten" is kept. This works in all templates — both the default cleanup and custom templates.

Common correction phrases that are recognized: "no wait", "I meant", "actually rather", "correction", and similar patterns in German and English.

Voice Commands

Verba recognizes spoken formatting commands and converts them to actual formatting. Say "new paragraph" to insert a paragraph break, "comma" to insert a comma, or "bullet point" to start a list item.

Supported commands:

Command	Result
"New paragraph" / "Neuer Absatz"	Paragraph break
"New line" / "Neue Zeile"	Line break
"Period" / "Punkt"	.
"Comma" / "Komma"	,
"Colon" / "Doppelpunkt"	:
"Semicolon" / "Semikolon"	;
"Question mark" / "Fragezeichen"	?
"Exclamation mark" / "Ausrufezeichen"	!
"Bullet point" / "Aufzaehlung"	- (list item)
"Number one/two/three" / "Nummer eins/zwei/drei"	1. / 2. / 3.

Commands work in any language — speak naturally and Verba will interpret them.