Skip to main content
This page tracks releases of the nebuly-platform Helm chart used to deploy Nebuly in your own infrastructure. Each entry lists the chart version and what changed. See Self-Hosting Nebuly for upgrade instructions.
​

​
Features

  • MCP: MCP answers can open the corresponding chart directly in Nebuly.

​
Bug fixes

  • Bug fixes and improvements.
​

​
Bug fixes

  • Bug fixes and improvements.
​

​
Bug fixes

  • Bug fixes and improvements.
​

​
Features

  • Model from tags: when traces carry no model name, a tag can be mapped as the model.

​
Bug fixes

  • Bug fixes and improvements.
​

​
Features

  • Topic cleanup: new optional reprocessing.cleanTopics job (disabled by default).

​
Bug fixes

  • Bug fixes and improvements.
​

​
Bug fixes

  • Bug fixes and improvements.
​

​
Features

  • Google Agent Platform: Gemini models (Vertex AI / Enterprise Agent Platform) can now be used for LLM tasks instead of OpenAI, via the new google values. openAi.enabled now defaults to false: set it explicitly if you use OpenAI.
  • ROI and AI Proficiency: activation moved from project to tenant level.

​
Bug fixes

  • Bug fixes and improvements.
​

​
Bug fixes

  • Bug fixes and improvements.
​

​
Bug fixes

  • Bug fixes and improvements.
​

​
Bug fixes

  • Bug fixes and improvements.
​

​
Bug fixes

  • Bug fixes and improvements.
​

​
Features

  • ROI: redesigned ROI page with drillable breakdown tables and empty states.
  • Navigation: new sidebar split into Monitor, Act and Reports.
  • AI Fluency Index (AIFI): the score now takes the interaction complexity tier into account.
  • Vertex AI: added as a supported LLM provider in traces.

​
Bug fixes

  • Bug fixes and improvements.
​

​
Bug fixes

  • Bug fixes and improvements.
​

​
Bug fixes

  • Bug fixes and improvements.
​

​
Bug fixes

  • Bug fixes and improvements.
​

​
Features

  • AI Spend: new page tracking AI expenditure across teams, tasks and models.
  • Model Savings: new page showing the model tier mix and potential savings from cheaper models.
  • New app shell: double sidebar with products, Nebuly pages and new AI Proficiency pages.
  • ROI: hourly cost can now be configured per taxonomy class or per tag.
  • Rolling window: time series charts support rolling windows.
  • Experience reports: reworked default reports for employee-agent projects (Topics, Team, Market, Tenure).
  • AIFI: multi-label classification with updated categories.
  • PII: improved PII detection.

​
Bug fixes

  • Bug fixes and improvements.
​

​
Features

  • LLM model and provider: inferred from traces when not sent explicitly.

​
Bug fixes

  • Bug fixes and improvements.
​

​
Features

  • Organization settings: always available; the frontend.enableOrganizationSettings value has been removed.
  • LLM model and provider: new interaction fields.
  • Langfuse v4: support for Langfuse v4 traces.

​
Bug fixes

  • Bug fixes and improvements.
​

​
Features

  • OpenAI models: deployments are now configured by tier (openAi.tier1ModelDeployment, tier2ModelDeployment, tier3ModelDeployment), replacing gpt4oDeployment, gpt5Deployment and translationDeployment.
  • Taxonomy steering: steering is now the default for category generation and extends to sub-categories.

​
Bug fixes

  • Bug fixes and improvements.
​

​
Features

  • Ingress: ingress templates refactored to work with any ingress controller.
  • AIFI: display names used in filters and group-bys.
  • Taxonomy: wildcard tenant supported in backend.taxonomyRoleConfig.

​
Bug fixes

  • Bug fixes and improvements.
​

​
Features

  • Taxonomy groups: classifier targeting (user query or assistant answer), retroactive application on a time window with estimate, cohort filters in class definitions and hierarchy-aware category moves.
  • User groups: composite filter logic (OR, nested conditions), new filters (active weeks, interactions per session) and rolling windows.
  • ROI: choose the taxonomy group used for time per task, with the reason shown for each estimate.
  • Global search: search interactions and reports across all visible projects.
  • Reports: new report header with inline editing; admins can reorder workspace reports.
  • Moderation and business issues have been removed.

​
Bug fixes

  • Bug fixes and improvements.
​

​
Features

  • Ingestion: new ingestionWorker.settings.skipProcessingForProjects value.

​
Bug fixes

  • Bug fixes and improvements.
​

​
Features

  • Taxonomy: new taxonomy creation UI and reserved categories.
  • Filters: improved filter UI; global filters can be saved as a new project.

​
Bug fixes

  • Bug fixes and improvements.
​

​
Features

  • AIFI: AIFI dimensions exposed and updated classes.

​
Bug fixes

  • Bug fixes and improvements.
​

​
Bug fixes

  • Bug fixes and improvements.
​

​
Bug fixes

  • Bug fixes and improvements.
​

​
Features

  • Taxonomy: custom groups shown and editable from interaction details.
  • User groups: users can be grouped via tags.
  • Cumulative cost: new metric.

​
Bug fixes

  • Bug fixes and improvements.
​

​
Features

  • OpenAI: default openAi.apiVersion updated to 2025-03-01-preview.

​
Bug fixes

  • Bug fixes and improvements.
​

​
Bug fixes

  • Bug fixes and improvements.
​

​
Features

  • Taxonomy: new backend.taxonomyRoleConfig value to set the minimum role allowed to edit the taxonomy.

​
Bug fixes

  • Bug fixes and improvements.
​

​
Features

  • Taxonomy: reachable directly from reports.

​
Bug fixes

  • Bug fixes and improvements.
​

​
Features

  • Monitoring: ClickHouse metrics collected by the OpenTelemetry collector.

​
Bug fixes

  • Bug fixes and improvements.
​

​
Features

  • Taxonomy: moved into Settings.
  • User groups: manual refresh, plus automatic daily refresh for all groups.
  • Sidebar: favorite and visibility toggles for reports.
  • Topics report: rate metrics added.

​
Bug fixes

  • Bug fixes and improvements.
​

​
Bug fixes

  • Bug fixes and improvements.
​

​
Features

  • Deep links: simpler links to conversations and interactions (/go/conversations/<id>, /go/interactions/<id>).

​
Bug fixes

  • Bug fixes and improvements.
​

​
Features

  • Taxonomy: multi-label custom groups, cleanup tab in Data management and role-based access.
  • ROI: various improvements.
  • Onboarding: updated product tour.

​
Bug fixes

  • Bug fixes and improvements.
​

​
Bug fixes

  • Bug fixes and improvements.
​

​
Features

  • GPT-5: new openAi.gpt5Deployment value.

​
Bug fixes

  • Bug fixes and improvements.
​

​
Bug fixes

  • Bug fixes and improvements.
​

​
Features

  • ROI: measure the business impact of AI directly in Nebuly. The new ROI Report surfaces hours and value saved, and ROI Settings let you configure employee cost and time-per-action estimates. Available in beta on both v1 and v2 — enable it from Settings > Feature Flags > ROI Analysis.
  • Frontend V2: the new frontend can now be enabled for self-hosted deployments. Admins are notified in-app through a new upgrade panel when a newer frontend version is available.
  • AI Fluency Index (AIFI): various feedback-driven improvements to the AIFI score.