mirror of
https://github.com/mastra-ai/mastra.git
synced 2026-11-17 08:44:46 +00:00
feat(@mastra/memory): use markdown formatting instead of XML inside working memory tags (#3396)
While making a demo for a user I noticed working memory updates were alot faster if I told the agent to format inner text in MD instead of XML. This makes sense since it doesn't need to print all the extra tags/tokens. I also talked with a user who was confused and thought the xml working memory data was meant to be stored in a structured format for db access/filtering (tags as columns or something like that). Using MD instead should make it more obvious that this is unstructured text and make it less confusing.
This commit is contained in:
parent
5f169dcb89
commit
3e72f94cde
9 changed files with 240 additions and 107 deletions
Loading…
Add table
Add a link
Reference in a new issue