I’m using the Twitter block to fetch tweets, but I’m running into issues with irrelevant and low-quality content, such as:
- Tweets in other languages (even though I want only English),
- Scammy content, especially related to crypto/bitcoin,
- Unwanted retweets and replies.
I tried adding search operators directly into the query string like this:
lang:en -is:retweet -is:reply -bitcoin -crypto
But unfortunately, this doesn’t seem to have any effect — the block still returns tweets outside these filters.
It would be really helpful if:
- These operators could be properly parsed and passed through the API,
- Or if the block included dedicated filter fields for language, content type (e.g., original tweets only), and keyword exclusions.
Is this something you could consider for an update? It would greatly improve data quality for more refined use cases.
Thanks so much in advance!