01. Voice profiles

Posts that sound like the founder. Or the engineer. Or the company.

The reason most AI-drafted posts read like AI is that the model has no idea who the post is supposed to sound like. Squawk solves that with voice profiles. Each profile captures tone, recurring topics, sample posts, and explicit anti-patterns. Drafts pull from that profile so the post reads like a person, not a brand template.

Three bluebirds in differently colored aviation headsets, each broadcasting in their own voice
Screenshot todoVoice profile editor: tone, topics, examples, anti-patternsapps/marketing/public/images/features/voice-profiles.png

What it does

The capabilities, one at a time.

Tone, topics, examples, anti-patterns

Each profile captures how the person actually writes (tone), what they care about (topics), how good output looks (examples), and what to avoid (anti-patterns).

Profiles per teammate and per brand

Multiple profiles in one workspace. The company can have its own voice. Each teammate can have their own. Drafts pick the right one for the channel and account.

Sample-driven calibration

Paste in a few real posts the person has already written. The profile uses them as anchors so drafts converge on their actual cadence.

Profile gets sharper over time

When you edit or reject a draft, Squawk records the change as a signal. Later drafts incorporate it. The profile improves as you use it.

Screenshot todoWorkspace view: company profile + teammate profilesapps/marketing/public/images/features/voice-profiles-list.png

Why it matters

The reason this exists.

Posts that come from real people outperform posts that come from a brand handle. The hard part is doing that at scale: each person needs their own voice, their own approval flow, their own connected accounts.

Squawk's voice profiles are the layer that makes multi-author content feel like a small army of distinct voices instead of one corporate megaphone with different headshots on top.