hey! just wanted to share an update, a while back I posted about a tool I was building to generate synthetic datasets. I had said I\u2019d share it in 2\u20133 days, but ran into a few hiccups, so sorry for the delay. finally got a working version now!<\/p>\n
right now you can:<\/p>\n
- \n
- give a query describing the kind of dataset you want<\/li>\n
- it suggests a schema (you can fully edit \u2014 add\/remove fields, tweak descriptions, etc.)<\/li>\n
- it shows a list of related subtopics (also editable \u2014 you can add, remove, or even nest subtopics)<\/li>\n
- generate up to 30 sample rows per subtopic<\/li>\n
- download everything when you\u2019re done<\/li>\n<\/ul>\n
there\u2019s also another section I\u2019ve built (not open yet \u2014 it works, just a bit resource-heavy and I\u2019m still refining the deep research approach):<\/p>\n
- \n
- upload a file (like a PDF or doc) \u2014 it generates an editable schema based on the content, then builds a dataset from it<\/li>\n
- paste a link \u2014 it analyzes the page, suggests a schema, and creates data around it<\/li>\n
- choose \u201cdeep research\u201d mode \u2014 it searches the internet for relevant information, builds a schema, and then forms a dataset based on what it finds<\/li>\n
- there\u2019s also a basic documentation feature that gives you a short write-up explaining the generated dataset<\/li>\n<\/ul>\n
this part\u2019s closed for now, but I\u2019d really love to chat and understand what kind of data stuff you\u2019re working on \u2014 helps me improve things and get a better sense of the space.<\/p>\n
you can book a quick chat via Calendly, or just DM me here if that\u2019s easier. once we talk, I\u2019ll open up access to this part also<\/p>\n
try it here:<\/strong> datalore.ai<\/a><\/p>\n<\/div>\n
submitted by \/u\/Interesting-Area6418 <\/a>
[link]<\/a><\/span> [comments]<\/a><\/span><\/p>
