← All work

Content creators & social media managers

ContentOS

An AI content operating system that turns a creator's own performance data, plus any external source, into a full calendar of platform-ready drafts.

In daily use, built for a working creator2026Next.jsTypeScriptSupabaseClaude APIApifyFullCalendarTailwind CSS

The problem

Creators run on instinct. The data about what actually works sits buried in platform dashboards nobody opens: which hooks hold attention, which topics land, when the audience is online.

Meanwhile every draft starts from a blank page. Ideas live in one app, analytics in another, the calendar in a third, and none of them talk to each other.

The insight

Every AI writing tool starts from a blank prompt. That is the mistake. A creator's own analytics already answer the question 'what should I post': which hooks held attention, which topics landed, when the audience was actually awake. But that data sits in a dashboard nobody opens, and the writing tool never sees it.

So the leverage isn't generation, it's grounding: making the performance history the prompt. Once analytics are the input, drafts stop sounding like AI and start sounding like the creator's best week.

What it does

A month of drafts grounded in what the audience already proved it wants, instead of a blank page every morning.

Analytics intelligence

Pulls a creator's full TikTok post history (views, likes, hashtags, transcripts) and identifies the top-performing hooks, topics, and posting windows.

Source ingestion

Feed it anything: an article URL, a YouTube video, a PDF, raw text. It extracts the substance and cross-references it against what the audience already responds to.

Voice-calibrated drafting

Generates platform-adapted drafts calibrated to the creator's own voice examples, not generic AI copy.

Content calendar

Approved drafts land on a monthly calendar with platform-colored scheduling. Idea to planned post in one flow.

How it works

  1. Scrape

    Apify pulls the creator's full post history with engagement data and transcripts.

  2. Analyze

    Claude identifies top hooks, best topics, posting windows, and audience signals. Insights persist to Supabase.

  3. Ingest

    External sources (URL, YouTube, PDF, text) are parsed and stored as idea fuel.

  4. Generate

    Ideas become per-platform drafts, calibrated against voice examples and the analytics layer.

  5. Schedule

    Approved drafts drop onto the calendar as scheduled content pieces.

Who this is for

A fit if

  • Creators and social media managers posting several times a week across two or more platforms.
  • Anyone running content on someone else's behalf (a founder's channel, a personal brand, a small roster) where voice consistency actually matters.
  • Operators with a back catalogue substantial enough to learn from.

Probably not for you if

It isn't for occasional posters. With a thin post history there is nothing for the analytics layer to learn from, and you would be paying for a very expensive text generator.

What this would take

A system like ContentOS is a Build engagement: one custom workflow, deployed in your stack, in four to six weeks. Most teams start with an Audit to confirm content operations is genuinely the bottleneck worth automating.

Audit

2 weeks

Build

4–6 weeks

Retainer

per month · 6-month minimum

Full detail →

Work

Site

Contact

myles1908@hotmail.com

The fastest way to reach me. I read every email myself.

© 2026 Myles Smith. All rights reserved.

Designed & built by Myles Smith