Manifesto
A brand has one voice.
Holding it is the hard part.
Why Plume exists
A brand speaks with one voice for as long as one person writes. The moment there are two, then ten, then models that remember nothing, the voice scatters. Never all at once: a post slightly too casual, a press release slightly too stiff, and six months later the brand no longer recognises what it publishes.
Generic tools cannot help, because they hold no memory of the brand. Every generation starts from nothing. Consistency then falls back on one person — the one who rereads everything, rewrites it, and becomes, without ever choosing to, the single point of failure of the whole editorial chain.
Plume exists so that a brand's voice stops being a discipline one person carries, and becomes an asset that compounds.
Where we are going
Toward a world where brand voice is a layer of the information system: defined once, verifiable, and depended on by everything else — the way a database or a design system is. An operating system for voice, not one more text generator.
What that commits us to
Content expresses an identity
It is not output to be produced by the yard. Every piece is a moment where the brand states itself — which is why Plume keeps the memory instead of starting over each time.
Infer rather than ask
A tool that asks fifteen questions before writing a line charges the writer for the work it was meant to spare them. Plume infers what it can from the brand, and asks only what it cannot know.
Proof, not a promise
“In your voice” stays unverifiable until someone measures it. Every piece ships with a match score, across named dimensions, that you can challenge.
The last signature is human
Plume writes, proposes and measures. It never publishes for you unless you asked it to: a brand commits the people who carry it, not a model.
If your brand writes a lot and rereads itself badly, that is exactly the problem we took on.
Start for free