Overwhelmed by the constant flood of news?
SnapNews cuts through the noise, delivering only the most relevant, personalised information directly to user's email.
It's a combination of
- Google Search API
- GPT-4o mini API
- Perplexity API
User enters
- Topic of your interest (can be a prompt)
- Interval how often you want to receive summarised news
Steps my AI engine takes:
- GPT, understand Topic (prompt) and generate 3 search queries for Google API
- Google Search gets 10 results for each query within given Interval
- Google Search also gets "old news" for those queries to exclude them from summarisation
- GPT reviews all results, filters out old news, validates which results are relevant for given Topic
- GPT passes filtered links to Perplexity which opens them and creates Newsletter
- final Newsletter is sent to users email address
Tell me what do you think about this. Good, bad, potential improvements, anything that could help me improve it.