EH Lab

checking...
Mode ?
First query loads the model (~7 s)
Top K ?
?

Expected record format — one JSON object per line

{
  "message_text_translated": "The main text content to index",
  "author":    "optional",
  "channel":   "optional",
  "posted_at": "optional ISO timestamp",
  "tags":      ["optional", "array"],
  "url":       "optional source URL"
}
Vectors ?
Shards ?
Index size ?
MB
Re-encode / Config ?

This will permanently delete all encoded vectors. Upload files are preserved. Continue?