fix: strip raw provider payloads from task result endpoint

fix: address review serialization and url encoding

dev: Humanize the reason labels

dev: Adding per-domain selection for feddback

dev: Apply feedback to multi-domain

dev: add hitl quick confirmation actions

dev: Refining Domain searches

feat: request-time generated preference forms

feat: three-layer user profile with domain-scoped context fusion

feat: surface generic recommendation items in API and UI

fix: persist recommendation result when task hands a result model (feedback 400).

refactor: centralize id generation

fix: attach stable feedback ids to recommendation messages

dev: Let chat become a 1s pooling streaming

dev: letting feedback status persisted into conversation status

dev: Feedback added to recommendation results

dev: additional user-kind filter in CMS

DEV: Debug -> Admin Dashboard with stats and user mgnt

clean: Removed legacy no-caller functions/modules

dev: User role-gated Debug page access

hardening: unifying recommendation task IDs

fix: persist preferences in user profile

frontend: Display message and input style

dev: IntroduceSQLAlchemy + Alembic for DB versioning and ORM

dev: Conversation branch selection state stored

dev: Making stream endpoint graph aware

dev: Persist collect-confirm HITL state

dev: Service type are aet Auto as we have intent detection

dev: Chat History are made Tree-shaped structure for editing support

dev:frontend: Page allowance for user message editing

test: frontend contract generation

API Testing playground implemented

Refactored css styles to separate folders and minor fixes

Frontend debug page

fixed some bug in processing logic

FIRST VERSION with real agent!!

fixed the preference renew; and shorten the prompt to save tokens

[major updates] add user profile and add LLM responds

changed map ui

update back and front end

Useful settings for huggingface space deployment

jn's part - a base of front&back-end, complete the basic interaction requirements