how avendo works.

getting started

1. sign in with your email. a magic link or a six-digit code, no password.
2. paste a product link: your shop page, website, App Store or Google Play listing. avendo reads the page and builds a brand kit: name, voice, value props and images it may use.
3. research starts on its own. avendo reads the sources that are switched on for your workspace and groups what people say into angles.
4. a first week of drafts lands in Posts. swipe right to post, left to skip.
5. connect your accounts in Settings → Accounts so swiped posts have somewhere to go. the setup check lists anything missing.

no footage? three clear phone photos of the product are enough to start. you can also send customers a link to send you short clips, with their consent stored next to each clip.

research sources

avendo keeps the words and drops the people. at intake, commenter names, handles, avatars and profile links are removed. each quote keeps its text, rating, date, source page and source kind.

angles, hooks, scripts

an angle is one reason people care, grouped from many quotes: a pain, a desire, an objection, an outcome or a use case. each angle shows its real quotes and a strength score. hooks are opening lines for that angle, ranked. scripts are short UGC-style scripts in beats: hook, problem, proof, call to action, written for a slider stop.

research shapes the wording of a post. it is never shown as a quote from a real customer.

steal this format

blueprints are post structures you can apply to any angle. they borrow the shape, never anyone's content.

• the honest demo
• pov: you finally…
• 3 things nobody tells you
• myth vs fact
• before / after
• the question everyone asks
• the 5-second unbox
• swipe-through how-to
• one big line

rendering & AI labels

“make post” turns a script into one of three formats:

• UGC video: 9:16, your brand-kit footage and photos with handheld drift, push-ins and pans, an AI voiceover and word-by-word captions burned in.
• carousel: 4:5 or 1:1 slides.
• static: one image.

labels you can't switch off

any synthetic element (an AI voice, an AI presenter, AI-generated imagery) means an AI label is burned into the media and the platform's own AI-content flag is set when posting. platform AI and branded-content labels are never removed. every render writes a receipt listing its inputs, providers, synthetic elements and credit cost.

presenters

AI presenters are off by default. when they're switched on for a workspace, a presenter needs a persona with a consent record: either a real person's signed likeness consent, or a licensed synthetic avatar. a presenter is never presented as a real customer.

changes

there's no timeline editor. ask for changes in plain words on the draft (“shorter”, “more casual”, “show the price”) or move the style slider; avendo rewrites and re-renders.

posting & accounts

connect as many accounts as you own on Instagram (business or creator), TikTok, YouTube and Facebook Pages. each account has its own type, posts-per-day limit, posting times and time zone.

• recommended (default): avendo picks the accounts where a post should do best and says why.
• all: every account that fits the format, each with its own version and time.
• pick: tick accounts yourself and save selections like “launch” or “everyday”.

the same post from several accounts gets flagged as spam, so each account gets a different hook, caption or cut, and linked accounts on one platform post hours apart. your accounts never like, comment on or share each other's posts. only real accounts you own.

the style slider

five stops from native (reads like a normal post) through soft, balanced and direct to ad (a clear offer). it changes tone, hook, how prominent the product is and the call to action. it never changes who the post is from.

autopilot levels

you also set posts per day, a weekly render budget, trend drafts per day and the hours between linked accounts.

testing

• hook tests: a post goes out with three openings on the same body, as trial posts or spaced posts. the winner keeps going.
• fix and re-run: when viewers leave at a specific second, avendo makes a V2 (new hook, shorter scene or tighter ending) and renders it.
• refresh: when a winning angle starts to fade, it comes back in a new format.

early numbers are marked provisional and never trigger automatic changes. a weekly digest sums up what posted, what won and what it cost.

replies

avendo drafts replies to comments and messages on posts it published, and flags complaints. nothing is sent until you approve it. reply permissions are waiting for platform review on some networks, so availability depends on the account.

Yieldline

Yieldline is a separate product for paid ads. avendo never spends money, creates ads or edits budgets. when an organic post beats the account's usual 3-second rate and retention, you can offer it to Yieldline as a promotion candidate, or let full autopilot do that. it arrives in Yieldline as a draft; Yieldline decides whether to spend, within its own approvals and limits.

to link the two, open Settings → Yieldline in avendo, enter your Yieldline project and workspace IDs, and paste the one-time inbound token into Yieldline. Yieldline can then send paid results and briefs for new versions back to avendo.

MCP

avendo runs an MCP server at https://avendo.io/mcp (streamable HTTP). create a token in Settings → MCP, then add it to your client:

{
  "mcpServers": {
    "avendo": {
      "url": "https://avendo.io/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_AVENDO_MCP_TOKEN"
      }
    }
  }
}

reading is free. anything that posts or uses credits returns a preview and asks for confirmation first. paid promotion is not an avendo tool; it goes through Yieldline's own MCP and approvals.

credits

a research pull costs 1/20 credit and a render ½ credit. plan credits reset every period; top-up credits never expire. see pricing.