Daily digest
Only items the LLM flagged as relevant. Everything else was auto-skipped.
Sources
Add any website. If it has an RSS feed it's used automatically; otherwise the page text is scraped.
| Name | Type | Last check | On |
|---|
Instructions to the LLM
This is your standing filter. Every article is judged against it — relevant items reach your digest, the rest are dropped.
Model
Cheap & fast for high volume.
Daily run hour (server time, 0–23)
Automatic run once a day at this hour.
Max items per source per run
Caps cost. 15 is plenty for a daily digest.
Outbound proxy (optional — leave blank)
Only fill this if a source starts blocking the server. Format: http://user:pass@host:port. Applies to scraping only, not the LLM.
Monitoring
LLM requests sent and their estimated cost. Each scraped item = one request.
By project
| Project | Requests | Tokens | Cost |
|---|
By model
| Model | Requests | Tokens | Cost |
|---|
Leads
Paste company URLs (one per line). Each site is crawled for CRM/sales signals, contacts are extracted, and Gemini scores ICP fit 0–100. Higher = hotter lead.