How to dictate a JSON config object with quoted keys
JSON requires quotes around every key, which is two extra spoken phrases per line compared with YAML. VoiceGem turns “open brace double quote camel case retry limit double quote colon 3 comma” into a correctly quoted and punctuated entry with nothing to clean up.
A fifteen-word JSON entry costs about 17 seconds typed against about 5.5 seconds spoken. (sources: Dhakal, Feit, Kristensson & Oulasvirta, “Observations on Typing from 136 Million Keystrokes”, CHI 2018, 168,000 participants, 136 million keystrokes; Yuan, Liberman & Cieri, “Towards an Integrated Understanding of Speaking Rate in Conversation”, Interspeech 2006, 2,438 American English telephone conversations (Switchboard corpus). The speech figure is conversational rate, not a measured dictation rate — no such measurement exists.)
{"retryLimit": 3,"logLevel":"debug"This tool runs entirely in your browser. Nothing you type is sent anywhere, stored, or logged.
Step by step
- 1
Open the object
Say “open brace”. The character keeps its spaces, which is what a JSON formatter produces anyway when the object spans lines.
- 2
Quote every key
Say “double quote”, the casing command and the words, then “double quote” again. JSON has no unquoted keys, so this pair is unavoidable on every line.
- 3
Say “colon” then the value
Speak the separator and the value. The colon attaches left, giving "key": value with the spacing JSON formatters use.
- 4
Say “comma” between entries
Speak the separator after each value except the last. JSON does not allow a trailing comma, and VoiceGem emits exactly what you say without checking.
JSON costs two more spoken phrases per line than YAML, and nothing else differs
Both formats dictate cleanly. The difference is quoting: every JSON key needs a pair of quote phrases that a YAML key does not, which on a twenty-key configuration file is forty extra phrases for exactly the same information.
That comparison is worth making before you start rather than after. VoiceGem handles both correctly and neither needs a cleanup pass, so if the choice of format is genuinely yours and you intend to dictate the file, YAML is measurably less to say.
Where doing it by hand breaks down
Editors auto-pair quotes and braces in JSON and will validate the document as you type, so a short object is quick and mistakes surface immediately.
Long configuration files are where the quoting cost compounds in both directions. Typing forty quote characters is forty shifted keystrokes, and every unbalanced one is a parse error rather than a formatting problem. Dictating the file in VoiceGem places each quote where you said it, and a missing one is visible in the demo output before it reaches your editor.
Doing this somewhere else? the browser-based speech-to-code converter, the same thing for dictating YAML configuration, how dictating Go doc comments works, or dictating Javadoc comments, step by step.
Common questions
No. Quotes, colons, commas and brackets all bind the way JSON formatters write them, so a dictated object is valid as it stands.
Two per key, for the opening and closing quote. On a twenty-key file that is forty extra spoken phrases for identical information.
No. The Deterministic Pass emits exactly what you say and never validates syntax, so your editor or parser catches it.
Dictate a two-key object and count the quote phrases you said.
Open the tool