v1.4.0
v1.4.0
This release introduces analytics & reporting, multi-language translation support, configurable agent disclaimers, LLM analyst context enrichments, and various stability improvements.
Analytics & Reporting
Added a full Analytics & Reporting module with Agent Performance and Node Analytics pages, including pre-aggregated daily stats with real-time incremental updates and a Celery batch fallback.
Added CSV, Excel, and PDF export for Agent Performance and Node Analytics reports.
Moved AI conversation metrics from /recordings/metrics to /analytics/metrics with added agent and date filtering.
Translations & Languages
Added multi-language translation support with full CRUD for languages and translation keys.
Conversation start now resolves translations (welcome message, title, disclaimer, possible queries, thinking phrases) based on the Accept-Language header.
Added agent info endpoint to retrieve available languages before starting a conversation.
Agent Disclaimers
Added configurable HTML disclaimer field (input_disclaimer_html) to agents, rendered below the chat input. Disclaimer is now server-driven, replacing the previous static frontend prop.
LLM Analyst Enrichments
Added configurable dynamic context enrichments for LLM analysts, allowing selection of data sources (Zendesk tickets, knowledge base content) and workflow node types to include in analysis context.
Conversations & Filtering
Enhanced conversation filters with agent, KPI score ranges (customer satisfaction, quality of service, resolution rate, efficiency), and empty conversation exclusion.
Added 4 new sort fields: customer satisfaction, quality of service, resolution rate, efficiency.
Stability & Fixes
- Reduced schema fetch calls with 5-minute global cache.
- Added
localhost:8022to CORS allowed origins. - Conversation Redis cache now cleared after finalization.
- Configurable subset of messages used for hostility analysis to reduce token usage.
- Updated subdomain handling in connection tester.
- Import cleanup and formatting improvements in ML model files.
Tooling & Packaging
- Added
rufffor Python linting andprettierfor frontend formatting. - Frontend project renamed to
genassist-frontend. - Chat plugin version bumped to
1.0.32.
New API Endpoints
| Method | Endpoint | Query Params | Description |
|---|---|---|---|
GET |
/analytics/agents |
agent_id, from_date, to_date |
Daily agent execution stats |
GET |
/analytics/agents/summary |
agent_id, from_date, to_date, compare |
Aggregated agent stats with optional period comparison |
GET |
/analytics/agents/export |
format (csv/xlsx/pdf), agent_id, from_date, to_date |
Export agent performance report |
GET |
/analytics/agents/{agent_id}/nodes/breakdown |
from_date, to_date |
Node type breakdown for a specific agent |
GET |
/analytics/nodes |
agent_id, node_type, from_date, to_date |
Daily node execution stats |
GET |
/analytics/nodes/export |
format (csv/xlsx/pdf), agent_id, node_type, from_date, to_date |
Export node analytics report |
GET |
/analytics/metrics |
from_date, to_date, agent_id, compare |
Aggregated conversation KPI metrics (replaces /recordings/metrics) |
GET |
/analytics/metrics/daily |
from_date, to_date, agent_id |
Daily KPI metric averages |
GET |
/translations/languages/all |
— | List all languages (including inactive) |
GET |
/translations/languages |
— | List active languages |
POST |
/translations/languages |
— | Create a language (code, name) |
PATCH |
/translations/languages/{language_id} |
— | Update a language (name, is_active) |
DELETE |
/translations/languages/{language_id} |
— | Delete a language |
GET |
/translations |
— | List all translations |
GET |
/translations/{key} |
— | Get a translation by key |
POST |
/translations |
— | Create a translation (key, default, translations) |
PATCH |
/translations/{key} |
— | Update a translation (default, translations) |
DELETE |
/translations/{key} |
— | Delete a translation |
GET |
/conversations/in-progress/agent-info |
— | Get agent metadata and available languages |
GET |
/llm-analysts/available-enrichments |
— | Available context enrichments (filtered by tenant config) |
GET |
/llm-analysts/available-node-types |
— | Available workflow node types with labels |
Modified API Endpoints
| Method | Endpoint | Change |
|---|---|---|
GET |
/conversations/filter |
10 new query params: agent_id, customer_satisfaction_min/max, quality_of_service_min/max, resolution_rate_min/max, efficiency_min/max, exclude_empty. 4 new order_by values. |
GET |
/conversations/filter/count |
Same new filter params as above. |
Removed API Endpoints
| Method | Endpoint | Replacement |
|---|---|---|
GET |
/recordings/metrics |
GET /analytics/metrics (with additional agent_id, from_date, to_date, compare params) |
Full Changelog: v1.3.11...v1.4.0