Brand & Voice

Each workspace has a Brand & Voice profile that’s injected into every buddy’s system prompt as the leading block. Tuning it takes effect on the next chat turn. No deploy, no re-login.

Open Settings → Brand & Voice to edit.

Business profile

Field	Limit	Notes
`businessName`	200 chars
`website`	300 chars
`industry`	enum	`saas`, `ecommerce`, `fintech`, `healthcare`, `education`, `media`, `marketplace`, `agency`, `devtools`, `consumer`, `hardware`, `nonprofit`, `other`
`industryOther`	120 chars	Free-form; shown only when `industry` is `other`
`description`	2000 chars	One- to two-paragraph pitch the agent leans on when writing outbound
`audience`	1000 chars	Who you sell to; shapes who Prospect targets by default

Voice sliders

Four 1-to-5 sliders shape output tone. Defaults to 3 across the board (neutral).

Slider	1	5
Formality	Conversational	Executive-register
Humor	Dry	Playful
Respect	Peer-to-peer	High-deference
Energy	Measured	High-energy

Edits write through PUT /api/v1/brand-context. Voice values outside 1–5 are clamped; missing values fall back to 3.

Enrichment source

enrichedSource is one of manual (you edited the profile) or auto-detect (Waypath scraped website to populate defaults). Auto-detect stamps enrichedAt so you can see when the profile last re-synced.

Endpoints


GET    /api/v1/brand-context                  # current profile (or {})
PUT    /api/v1/brand-context                  # upsert (shallow-merges voice + links)
DELETE /api/v1/brand-context                  # wipe

All require auth + workspace context. Persisted per-workspace in ws_{id}_brand_context.json with the same encryption policy as integration configs.

How it reaches the agent

Server-side, the system prompt is assembled in order:

Brand block: built from this profile via getBrandContext()
Buddy persona intro
Base system prompt
Playbook block

Because the brand block is first, a voice-slider nudge takes effect on the next request. No restart required.