docs
everything to get the most out of brandmint: from your first swipe to the api, plus legal and data terms.
introduction
brandmint learns how you write and turns your ideas into finished posts you approve with a swipe. it reads your public posts to build a voice profile, then drafts in that voice, getting sharper every time you keep or pass on a draft.
there are two ways in: give minty a topic, or let it suggest ideas. both end in the same place: a stack of drafts you swipe through and schedule.
quickstart
- connect a source. paste your linkedin, site, or any public profile.
- let minty read. it extracts your voice, angles, and signature phrases.
- swipe. keep the drafts that sound like you, pass on the rest.
- schedule. kept drafts land in your calendar. drag to a day and set a time.
how voice learning works
every swipe is a signal. keeps reinforce patterns; passes (and the reason you give) tell minty what to avoid. over time your profile captures tone, structure, hooks, and topics, stored in memory, which you can view or wipe anytime.
swiping & preferences
the swipe deck is the heart of brandmint. drag right to keep, left to pass, or use the buttons. on a pass, minty asks a one-word reason so the next batch improves.
generating drafts
pick a platform; brandmint chooses the format and angle from your profile and returns five drafts. each carries informational label chips (tone, format) so you can see its reasoning.
calendar & scheduling
saved drafts appear as unscheduled in your calendar. drag one onto a day or open it to set a date and time. week and month views are available; scheduling wires to publishing via connectors.
connectors
connect linkedin, instagram, x, and more to publish directly and give minty more context. tokens are revocable at any time from settings.
api reference
authenticate with an api key from settings → api keys. pass it as Authorization: Bearer bm_... or X-API-Key: bm_.... all endpoints return json.
extract voice profile
analyzes your posts to extract tone, hooks, rhythm, formatting, vocabulary, themes, structures, and engagement patterns.
generate post
writes a linkedin post in your voice. if voice_profile is omitted, extracts it first.
check quality
scans a draft for ai tells, voice alignment, and factual accuracy. returns a cleaned version.
generate ideas
turns a seed topic into 4-6 post ideas with hooks, angles, and subtopics.
search posts
search your post history. supports semantic, keyword (bm25), hybrid, engagement, and recency ranking.
content strategy
analyzes your content performance and produces data-backed strategy recommendations.
profile
read or update your creator profile, content themes, and positioning.
save draft
save a finished post with an auto-generated share link.
rate limits
default: 1,000 requests per day per key. the rate limit resets at midnight utc. if exceeded, the api returns 429.
mcp server
brandmint exposes a hosted mcp endpoint that works with claude code, cursor, windsurf, and any mcp-compatible client. one line of config, all tools appear.
claude code
add to your .claude/settings.json or project config:
available tools
extract_voice— analyze posts, return 8-dimension voice profilegenerate_post— write a post in your voice (auto-extracts voice if needed)check_quality— scan for ai tells and voice alignmentgenerate_ideas— 4-6 ideas from a seed topiccontent_strategy— data-backed content performance analysissearch_posts— search post history (semantic, keyword, hybrid)get_profile— creator profile, stats, preferencessave_draft— save a draft with share link
webhooks
subscribe to events to keep your systems in sync.
draft.kept— a draft was swiped rightpost.scheduled— a draft was placed on the calendarpost.published— a post went live via a connector
legal
terms of service
by using brandmint you agree to use the product lawfully and retain ownership of the content you create. brandmint provides drafting tools; you are responsible for what you publish. accounts may be suspended for abuse or attempts to scrape others' private data.
privacy policy
we collect the public content you point us to, the swipes and edits you make, and account details. this trains your voice profile only. we never sell your data or use it to train models for other users.
data & security
- connector tokens are encrypted at rest and revocable anytime.
- your voice profile and memory can be exported or deleted on request.
- we read only public posts until you explicitly connect an account.