Decision first, copy second
The system answers what to make and why. Next to every recommendation sit the scoring factors, the sources, the risks and the confidence.
Knows what content a brand should make now and why
Every day, for every brand, a list of opportunities with the reason, the deadline and a ready brief. Once published, the system measures the result and sharpens the next recommendation.
Content is assembled across six tabs
The idea comes from one service, the copy from another, the plan lives in a spreadsheet, approval happens in a messenger, and the numbers sit in the platform dashboard.
- 01Idea and copyChatGPT
- 02VisualsCanva
- 03PlanNotion or a spreadsheet
- 04ApprovalTelegram or Slack
- 05PublishingThe platform app
- 06ReportingAn analytics service
A month later nobody can say why that topic was chosen, which version was approved and what actually worked.
How it works
Five steps from signal to conclusion. Then the cycle repeats, and every loop is sharper than the last.
- 01
Brand profile
You give a link to the site and social accounts. The system collects the product, the audience and the tone of voice, then asks you to confirm it.
- Goals, segments and content directions
- Banned topics work as a hard filter
- The profile is versioned and each recommendation remembers its version
- 02
Opportunities every day
Signals from external trends, audience questions, brand history, business events and strategy gaps are run through the brand profile.
- The reason and the scoring factors
- How long the topic stays alive
- Confidence and risks on a separate line
- 03
A brief, not a post
Several angles, each with its own archetype: explanation, checklist, case, demonstration, reaction.
- Goal, hook, key points and call to action
- Visual plan
- Links to the sources behind the facts
- 04
Approval by link
The client opens a link without signing up, comments on a specific block and makes the decision.
- Versions never get mixed up
- Who approved and when stays in the history
- An approver does not take a paid seat
- 05
Result and conclusion
The result is compared not against a market average but against the norm of this account for this format.
- Editor, audience and business reactions kept apart
- A weekly review with a conclusion for the week ahead
- A rejection counts just as much as a publication
Formats the system writes briefs for
Product close-up Reels, outfit of the day, food styling. The brief covers the hook, the shot list and the call to action; your team does the filming.
Videos are stock examples of the formats from Pexels, not client work.
Why this is not another post generator
Generating text stopped being an advantage long ago. The advantage is what happens before it and after it.
Relevance beats virality
A strong trend can be a poor fit for a brand, and then it never reaches the list. Banned topics are cut before scoring, not after.
Math scores, AI writes
Growth, velocity and author diversity are computed by a transparent formula. The AI takes those numbers and assembles the brief, but does not score anything itself.
Rejections are remembered too
The whole decision is stored: what was shown, what was chosen, what was rejected and why. Otherwise the system only learns from what got published.
Pay per brand, not per person
The billing unit is one active brand per month. Approvers on the client side never take a paid seat.
Solo
A freelancer with a few clients
$39per month
- Brands
- up to 5
- Users
- 1
- Roles and permissions
- no
- Reports
- basic
- Integrations
- no
- Success metrics
- standard
Agency
An agency with a team
$129per month
- Brands
- up to 20
- Users
- team
- Roles and permissions
- yes
- Reports
- client ready
- Integrations
- no
- Success metrics
- standard
Agency Pro
Many brands and your own brand on the reports
$299per month
- Brands
- over 20
- Users
- team
- Roles and permissions
- yes
- Reports
- your logo
- Integrations
- CRM
- Success metrics
- custom
Questions
No. The brief exports as a file or goes to a messenger, visuals are made where you already make them, publishing works the way it did. Integrations are connected step by step.
No. They open a link, see only their own material and make the decision. The link has a lifetime and can be revoked.
From open feeds, official platform interfaces, tracked competitors, comments under your own posts and whatever you add by hand. Grey scraping cannot be the foundation: it works today and the platform shuts it down tomorrow.
Banned topics and claims are cut before the score is computed. You can see which sources a recommendation is built on and how confident the system is. A rejection with a reason feeds into the next set.
Exactly as much as the data is complete, and the interface shows that. A value typed by hand is marked differently from one pulled off the platform. If one metric is missing, the rest of the conclusion is not thrown away.
Tokens are stored encrypted and only the minimum permissions are requested. Data from one workspace is not reachable from another. A connection can be deleted and the data exported. Training a shared model on your content is off by default.