feat: add conversation memory pipeline
This commit is contained in:
12
CHANGELOG.md
Normal file
12
CHANGELOG.md
Normal file
@@ -0,0 +1,12 @@
|
||||
# Changelog
|
||||
|
||||
## 0.1.1
|
||||
|
||||
- add an optional two-stage conversation memory pipeline with classifier and extractor models
|
||||
- store approved inbound and first-pass outbound memories back into the persona IdentityDB space with turn-trace metadata
|
||||
- skip re-persisting exact duplicate extracted memories by domain and statement
|
||||
- expose the new conversation memory pipeline through the public `conversation` module exports
|
||||
|
||||
## 0.1.0
|
||||
|
||||
- initial public BoxBrain framework release
|
||||
Reference in New Issue
Block a user