Run agent pipeline
Topic input
Agent pipeline
Phase 1 — Intelligence (parallel)
Marketing director
Research agent
SEO agent
Phase 2 — Production
Script agent
Creative director
Voiceover agent
Phase 3 — Intelligence
Analytics agent
Output
Final script
SEO brief
Marketing brief
Research brief
Creative notes
Pipeline stats
—
Total runs
—
Published
—
Voiced
—
Drafts
Pipeline
① Marketing + Research + Analyticsparallel
② SEO agentbriefed
③ Script agentfull brief
④ Creative Director reviewquality gate
⑤ Voiceover → MP3ElevenLabs
Topic ideas
AI-generated topic suggestions — refreshed daily, informed by analytics
Loading...
Video history
Topic / TitleStyleSEOCreative StatusDateActions
Loading...
Analytics & intelligence
—
Total
—
Published
—
Avg SEO
—
Avg creative
Analytics agent brief
Run the pipeline once to generate analytics.
Content gaps
No data yet.
Style breakdown
No data yet.
Recent activity
No data yet.
Settings
Worker URL

All agent calls route through your Worker. API keys are stored as Worker secrets — never here.

Settings saved.
Deploy commands
1. npm i -g wrangler && wrangler login
2. wrangler d1 create cycling-db → paste id into wrangler.toml
3. wrangler d1 execute cycling-db --file=schema.sql
4. wrangler kv:namespace create cycling-kv → paste id
5. wrangler secret put ANTHROPIC_API_KEY
6. wrangler secret put ELEVENLABS_API_KEY
7. wrangler deploy
8. wrangler pages deploy public/
Cron schedule

Auto-runs every Sunday at 10:00 GST (06:00 UTC).
Generates one script from the top analytics content gap.
Voiceover is skipped on scheduled runs to save credits — trigger it manually from History.

Change schedule in wrangler.toml → crons = ["0 6 * * 0"]