#thecompaniesapi extract 500+ domains for our dataset; keep monitoring but everything goes smooooth, claude haiku is actually better for our usecase than sonnet ; blasting just below the rate limits for hours until we have full dataset
#thecompaniesapi added model selection to our source system ; we can now run user queries on a model and our continuous extraction/datasets sources on another; became afraid of receiving a huge batch of user queries and make it create a huge Anthropic bill 😅 ; also merged my last 2 PRs to next production batch