How Audyense works: from your product to a post you approved

Audyense is built around one loop. You tell it about your product, it works out who your audience is, it writes for that audience, and it posts what you approve. Each step feeds the next, and each step can be revisited as the product changes. Here is what happens inside each one.

1. Knowledge: reading what you already have

Under Knowledge you add sources: specific URLs (your homepage, pricing page, a feature page, docs) and uploaded documents (a one-pager, a pitch deck, a spec, a PDF). Audyense fetches each one and reads it into an understanding, a structured account of what the product does, who it serves, how it is priced and how it describes itself.

It reads only what you add. It does not crawl the wider web for opinions about you, and it does not guess. If your pricing page says one thing and your deck says another, both end up in the understanding and you can see it. A scan costs one credit per source, and you can re-scan a source any time it changes.

2. Audience: who buys, and where they already are

From the combined understanding, Audyense builds your audience profile. The core of it is one to three ideal customer profiles, deliberately narrow: the role, the kind and stage of company, what triggers them to start looking, why the product fits and why it does not. Around each ICP it lists the places those people already spend attention: subreddits, YouTube channels, newsletters, podcasts, X accounts, the people they trust, and the phrases they search for.

Every named community and person is checked against a live search before you see it. A model asked for subreddits will happily invent one that sounds right; anything that does not resolve is dropped rather than shown. The profile also surfaces competitors that serve the same buyer, with positioning angles to test. It refreshes weekly, and a rebuild costs ten credits.

3. Content: one platform at a time

Content drafts are written for a specific platform, because a Reddit comment, a LinkedIn post and a YouTube script are different things and readers can tell when one has been reshaped into another. Audyense supports X, Reddit, Quora, LinkedIn, Facebook pages and groups, Slack and Discord communities, YouTube and TikTok scripts, and blog articles.

Each draft is grounded in your knowledge sources and aimed at your audience profile. It arrives with a short rationale explaining why this topic, for whom, and where. Short-form drafts cost one credit, long-form (scripts, articles) three. You approve, edit or reject. Feedback you leave on one draft shapes the ones that follow.

4. Posting: only what you approved

Under Settings you connect the accounts you post from. You authorise on each platform itself, so Audyense never sees a password, and you can revoke access there at any time. Approved drafts are posted from those accounts; nothing goes out without your approval. Posting costs one credit.

Around the loop: competitors and your website

Two workspaces sit alongside the loop. Competitors turns the audience research into decisions about positioning: who you are actually up against, angles to test, practical next actions. SEO scans your public website (up to 500 pages) for search fundamentals, answer structure for AI assistants, and machine readability, and turns the findings into a prioritised list of fixes that later scans verify. A scan costs twenty credits.

What it costs

Plans carry a monthly credit allowance and a number of connected accounts: Starter is 250 credits and 3 accounts, Growth 900 and 10, Scale 3,000 and 30. The free plan includes 100 welcome credits, enough to read your product and build a first audience profile, but no connected accounts. Extra credits come in packs and last a year. The allowance is always spent before anything you paid for.

A useful mental model: the subscription pays for keeping accounts connected, credits pay for the model and search work. Neither is charged for content you rejected.