## Process
1) Select group of wallets
- Script to output key context about the wallets, using some combination of wallet in context labels, affiliations, and text summarization of interests &affiliations

2) Search for relevant content, depending on newsletter use case
- What are people in our community doing
- What are job opportunities for people in our community to do
- What are media updates, in terms of articles, grants, or proposals that may be relevant to people in our community (and why)

3) Return articles + context about the article
- Author, title, date
- Text summary
- Collectors + WIC labels, Authors + WIC labels
- Referenced organizations + WIC labels

4) Post-process article further
- Query the affiliates of the articles, bring back response object with context, query GPT to ask why it would be relevant to the original set of wallets
-

5) Return final newsletter
- Prompt with use case, provide relevant information, collect output
- Further post-process, if required, for specific users

## Training data sources
- Newsletter styling
- Substack
- Medium
- Mirror
- Pallet (for jobs)
- Can scrape, but also have an email account that is subscribing to every Web3 newsletter I can find to continously source relevant data
-
- Twitter threads related to newsletters