product overview

hold your voice helps you write with ai without losing the way you sound. it builds a voice profile, checks drafts against it, and remembers which fixes you accept.

updated may 3, 2026 start here no images
quick answer

start here if you need the plain english map of the whole product before reading the feature pages.

use this for

  • explaining hold your voice to a teammate or an ai assistant
  • understanding how the app, api, mcp server, and byok plan fit together
  • deciding which wiki page to read next

steps

  1. create an account and finish onboarding.
  2. build or paste a voice profile.
  3. write or paste a document in the editor.
  4. run a voice check and review the highlighted signals.
  5. accept, dismiss, or rewrite only the lines that need work.

details to know

  • documents are the main unit of work. each document has a format, a voice score, a title, content, and optional sharing settings.
  • voice profiles hold the rules for how a brand or person should sound. team and byok plans can hold more than one profile.
  • the signal engine checks both ai-writing patterns and positive voice signals. the score is not a grammar score.
  • the public wiki is for product behavior. private admin tools, internal automation, and secrets do not belong here.

limits and edge cases

  • the product changes fast, so the wiki should be updated in the same commit as any user-facing behavior change.
  • public shared documents are not indexed. wiki pages are indexed.
ready to use the product instead of reading about it? open app