How Cadence works
The weekly rhythm, broken down.
Six stages. One loop. Every week your pipeline pulls what competitors just posted, turns it into a report, and hands you five ideas ready to move through a board.
You bring your own API keys.
Cadence is the software. Scraping (Apify), scripts + ideas (OpenRouter), embeddings (Google AI) all run on your keys encrypted with AES-256-GCM. You see every call, every cost. Rotate or revoke any time.
Scrape the recent window
Transcribe what's worth transcribing
A weekly strategist report
Generate ideas in your voice
- Novelty vs your past hooks — how different this is from what you've already posted.
- Voice match with your scripts — how closely the tone and cadence match your real posts.
Move, schedule, publish
And the loop closes.
The reels you actually posted become part of the next report's input. Hooks that did well get remixed forward. Ideas you rejected get dropped from future novelty scoring. Every run sharpens.
What Cadence isn't.
- Not:A video editorWe don't edit, upload, or host video. You edit where you already edit — we schedule what you're about to post.
- Not:A scheduler that posts for youInstagram's API doesn't make autoposting reliable for creators. Cadence lives on the planning side: what to make, when to make it, when it goes out.
- Not:A training dataset farmYour scripts, edits, and scraped data are yours. We don't train models on your content. We don't share it with third parties.
Your data stays yours
Tenant-isolated Postgres with row-level security. Your keys encrypted with AES-256-GCM.
Every cost visible
Each LLM run logs model, tokens, and cost. You see exactly what each report or idea batch costs.
BYOK, always
Apify, OpenRouter, Google AI — you connect the keys, you pay the providers. Cadence only charges for software.
Try it on three competitors.
Connect your keys, add three handles, generate your first report. Fifteen minutes end-to-end.
Start your workspace