onyx-dot-app · jmelahman · Aug 27, 2026
diff --git a/changelog.mdx b/changelog.mdx
@@ -6,6 +6,32 @@
 rss: true
 ---
 
+<Update label="v4.7.0" description="August 27th, 2026" tags={["New Features"]}>
+
+  ## v4.7.0
+
+  Onyx's interface is now available in Spanish, Portuguese, French, and German, alongside English.
+
+  Thank you to our contributors for helping make this release possible!
+
+  ## Key Improvements
+
+  ### Interface Language
+
+  You can now use the Onyx interface in **Español**, **Português**, **Français**, or **Deutsch**,
+  in addition to English.
+  Choose your preferred language from the new **Language** dropdown on the
+  [General](https://docs.onyx.app/overview/core_features/general_settings)
+  tab of your User Settings (`/app/settings/general`) — language names are always shown in their native spelling,
+  regardless of which language is currently active.
+
+  ---
+
+  Our [GitHub Releases](https://github.com/onyx-dot-app/onyx/releases?utm_source=docs&utm_content=changelog)
+  page has the full list of changes and bug fixes for each release.
+
+</Update>
+
 <Update label="v4.4.0" description="July 21st, 2026" tags={["New Features", "Deployment Changes", "Bug Fixes"]}>
 
   ## v4.4.0
@@ -24,7 +50,7 @@
  Providers](https://docs.onyx.app/deployment/authentication/sso_providers)** instead of environment variables.
  On the first upgraded startup,
  a one-time migration seeds your existing `AUTH_TYPE` / OAuth / OIDC configuration into a provider row that keeps the
  exact redirect and ACS URLs your identity provider already allowlists — no OAuth client changes are needed.
  **Keep the legacy variables in place through the upgrade** so the migration can read them.

  After upgrading, confirm the provider appears under **SSO Providers** and that logins still work,
@@ -41,7 +67,7 @@

  ### Hardened, non-root model server and web images

  The model server and web frontend now build on distroless, hardened images: no shell or package manager,
  and they run as a non-root user.

  <Note>
@@ -96,8 +122,8 @@

  ## Contributor Highlights

  - [@siriusfreak](https://github.com/siriusfreak) — after adding Tavily
    and Nebius Token Factory in v4.3.0 — contributed a full **LumApps connector** with checkpointed incremental indexing
    of published intranet content!
  - [@jack-watroba](https://github.com/jack-watroba) fixed MCP OAuth
    public-client PKCE registration in their first contribution!
@@ -140,7 +166,7 @@
  - Claude Sonnet 5 support with adaptive thinking; Fable 5 and Sonnet 5
    join the recommended models.
  - Vision and reasoning support for dynamic providers resolves from the
    LiteLLM cost map instead of hardcoded lists.
  - Recommended model lists update automatically from the OpenRouter
    catalog.
  - Gemini `embedding-2` (GA) available for cloud embedding.
@@ -176,7 +202,7 @@
  - Fixed the agent editor keeping rejected file uploads in state.
  - Fixed the empty OAuth credential modal for zero-kwargs OAuth
    connectors.
  - Fixed onboarding not scrolling at small viewports.

  ---

@@ -191,14 +217,14 @@

  SIEM-exportable audit logging, TLS and mutual TLS across the entire data plane, conversation-aware search scoping,
  hardened indexing for large files,
  and two new providers — Tavily and Nebius Token Factory — both from a first-time community contributor.
  Thank you to our contributors for helping make this release possible!

  ## Key Improvements

  ### Audit Logging

  Onyx now emits a normalized, SIEM-exportable audit-event stream — Splunk, Sentinel, Elastic, Chronicle,
  and AWS Security Lake can ingest it with no per-SIEM integration. Coverage spans authentication (login, logout,
  register, password reset, verification), admin configuration (LLM providers, connectors, credentials, API keys),
  access control (roles, user creation, group membership), and impersonation.
@@ -217,17 +243,17 @@

  Internal search now decides, per call,
  which connected sources to cover based on the agent's routing instructions and the current request.
  Tell an agent "check Zendesk first, then Confluence" and the search walks those sources in order,
  falling back to everything if needed. Scoped searches show it in the timeline:
  "Searching Zendesk" instead of "Searching internal documents."

  ### Hardened Indexing for Large Files

  Large spreadsheets are now streamed row-by-row through a file-backed pipeline instead of being rendered into memory,
  so an oversized workbook can no longer OOM an indexing worker. PDF text extraction runs in an isolated subprocess,
  so a pathological PDF can't take the worker down with it.

  ### New Providers: Tavily and Nebius Token Factory

  [Tavily](https://tavily.com) joins the web-search providers as both a search and content-extraction provider.
  [Nebius Token Factory](https://tokenfactory.nebius.com)
@@ -236,8 +262,8 @@

  ## Contributor Highlights

  - [@siriusfreak](https://github.com/siriusfreak) added both the Tavily
    and Nebius Token Factory providers in their first contributions!
  - [@greymoth-jp](https://github.com/greymoth-jp) fixed Enter during IME
    composition submitting list inputs — a long-standing annoyance for Japanese, Chinese, and Korean text entry!

@@ -246,7 +272,7 @@
  ### Security and Hardening

  - Connector read endpoints are restricted to curators and admins;
    previously any authenticated user could read connector configs.
  - Persona owner emails and internal MCP server URLs are hidden from
    basic users.
  - Curator group writes are scoped to the groups they actually curate.
@@ -271,22 +297,22 @@
    cleanup queue.
  - Repeated-error connector notifications reach every active admin
    instead of a single system inbox.
  - The admin stage-metrics view covers previously untimed pipeline
    stages, exposing lock-wait and contention.
  - Jira connectors without permission sync no longer require
    permission-reading credential scopes.
  - Notion hierarchies parent correctly when children index before their
    parents; standalone databases are no longer orphaned.
  - GitHub connector setup surfaces the actual validation error instead of
    a generic failure.
  - SCIM group-membership removals from Microsoft Entra ID now sync
    correctly.

  ### Deployment

  - API-server workers, connection limit, and threadpool size are
    configurable in Helm enabling concurrency headroom for clusters that lean on long-running agent features.
  - Model-server torch threads are capped to the cgroup CPU quota,
    eliminating CFS throttling on large hosts.
  - New Celery `clear_revoked` control command wipes the fleet-wide
    revoked-task set without restarts.
@@ -300,7 +326,7 @@
  - Fixed Confluence group sync not receiving scoped tokens.
  - Fixed custom branding not applying to the browser tab title and
    favicon on hard refresh.
  - DOCX files are no longer duplicated as plaintext now that the frontend
    renders DOCX natively.
  - Fixed the agent claiming it lacks tools after web search when Open URL
    is disabled.
@@ -373,8 +399,8 @@

  ### Indexing Memory and Reliability

  - PDF text extraction switched to pypdfium2 (with pypdf fallback),
    eliminating heartbeat-timeout deaths on large PDFs.
  - Google Doc parsing is size-bounded, fixing OOM crash-loops on tenants
    with very large docs.
  - Google Drive streams documents in sub-batches to bound connector
@@ -382,9 +408,9 @@

  ### Connectors

  - New Braintrust connector.
  - Google Drive follows shortcuts during indexing.
  - SharePoint supports targeted reindex of individual failed documents.
  - Slack: channel exclude list with optional regex, and permission sync
    scoped to configured channel filters.
  - Gong: configurable API base URL for non-US data residency.
@@ -419,7 +445,7 @@

  ### Notable Bug Fixes

  - Fixed unreachable Ollama and OpenAI-compatible model servers returning
    502 instead of a clear 400.
  - Fixed vision support detection for Bifrost, LiteLLM Proxy, and other
    dynamic providers, with a clearer error on oversized images.
@@ -485,7 +511,7 @@
  ## v4.1.0

  Personal Access Tokens gain user-selectable scopes, MCP servers from known providers can be connected via OAuth,
  voice moves to OpenAI's GA Realtime API, and six new community contributors land improvements across connectors, auth,
  and LLM providers. Thank you to our contributors for helping make this release possible!

  ## [Update] Deployment Changes
@@ -494,7 +520,7 @@
    If you build custom images or install additional packages on top of the Onyx image,
    verify compatibility before upgrading.
  - **The Helm chart now fails fast on 0.4.x → 0.5.x upgrades that would
    lose Vespa data**, protecting deployments that haven't completed the OpenSearch migration from accidental data loss.

  ## Key Improvements

@@ -510,9 +536,9 @@
  MCP servers from known providers can now be connected via OAuth directly from the authentication modal,
  with a guided connect-and-callback flow instead of manual credential entry.

  ### Voice on the GA Realtime API

  OpenAI streaming speech-to-text has been migrated from the deprecated Realtime Beta API to the GA Realtime API,
  putting voice mode on a supported foundation going forward.

  ## Contributor Highlights
@@ -521,14 +547,14 @@

  - [@ppomes](https://github.com/ppomes) fixed the ingestion API overriding
    document sources, corrected default temperature selection for the search tool,
    and optimized Google Drive indexing to skip unnecessary shared-drive lookups!
  - [@fejesd](https://github.com/fejesd) enabled prompt caching and sticky
    routing for OpenRouter and wired user identifiers into OpenRouter activity tracking!
  - [@amanda-nathan](https://github.com/amanda-nathan) added support for
    unauthenticated SMTP relays!
  - [@dimanjet](https://github.com/dimanjet) and
    [@mjakubowskimtt](https://github.com/mjakubowskimtt) both fixed ClickUp connector API URL handling!
  - [@psyray](https://github.com/psyray) fixed an nginx config typo in the
    no-letsencrypt compose file!

  And returning contributor [@beraterkanelcelik](https://github.com/beraterkanelcelik) made job timeouts,
@@ -544,12 +570,12 @@
  - S3 connector supports an optional AWS region for non-default partitions.
  - Confluence skips image attachments in the slim-doc pass when images are
    disabled, and permission sync handles draft-page 404s.
  - Jira Data Center v9 permission sync support.

  ### Auth and Security

  - CSV exports are sanitized against spreadsheet formula injection.
  - Login and signup use a native password field for proper browser
    integration.
  - Gmail addresses with dots can sign in via Google OAuth.

@@ -569,7 +595,7 @@
    heavy file usage.
  - Billing information and LLM provider listings are cached per tenant.
  - Frontend polling tuned: document sets only poll while syncing, and
    several SWR fetches were deduplicated.

  ### Deployment and Observability

@@ -599,18 +625,18 @@

  ## v4.0.0

  Vespa is gone, OpenSearch is now the only document index. We're shipping an opt-in Coding Agent,
  end-to-end visibility into indexing failures, permission sync inspection tooling, a public Search API,
  and Slack Enterprise Grid support.

  Thank you to our contributors for helping make this release possible!

  ## [Breaking] Vespa is Removed

  <Warning>
    **Do not upgrade to this release if you want to preserve your existing indexed documents but have not yet run the
    document index migration,
    which can be done from any version v3.x.** Vespa has been removed entirely in v4.0.0 and upgrading without first
    migrating will result in the loss of indexed documents and Project files,
    for which the only solution is re-indexing.

@@ -630,7 +656,7 @@

  ### Coding Agent (opt-in)

  v4.0.0 introduces a new Coding Agent — a sandboxed agent with file download and bash tool access,
  available as a standalone agent and as a tool that the regular chat agent can invoke.
  The Coding Agent ships **disabled by default**;
  admins can enable and configure it from a new section of the Admin Panel.
@@ -640,7 +666,7 @@

  Admins can now see exactly what failed during indexing and resolve it without rerunning entire connectors.
  Failed documents are surfaced with the underlying error,
  individual documents have a "Resolve" button that triggers a targeted reindex for just that document,
  and a "Resolve All" action handles bulk recovery. Indexing metrics are now always visible rather than gated,
  and the indexing errors modal supports pagination through large failure lists.

@@ -693,19 +719,19 @@
    submitting.
  - Typewriter streaming finishes faster after the LLM stream completes.
  - Tool calls can now access a `bash` tool (used by the Coding Agent).
  - All CSVs are passed to Code Interpreter for analysis.

  ### Connectors

  - **Slack Enterprise Grid** support.
  - HubSpot: search API for incremental polls and `batch_read` to eliminate N+1
    association fetches.
  - Confluence: Data Center websudo support, REST space permissions for 9.1,
    attachment downloads with non-scoped tokens, and pagination fixes.
  - Freshdesk: handle null `due_by` and paginate past 300 pages.
  - GitLab: fall back to username when GitLab masks the author name.
  - Asana: stop dropping teamless projects from explicit allowlists.
  - Jira: permission sync support for Data Center v10.

  ### Auth, Security, and Licensing

@@ -723,11 +749,11 @@

  ### Notable Bug Fixes

  - Fixed a tabular chunker crash on Google Sheets with multi-line cells.
  - Fixed the indexing pipeline not stopping after a document raised an error.
  - Fixed several DB-session-held-too-long issues in chat, indexing, deletion,
    and embedding paths.
  - Fixed asyncpg `statement_cache_size` regressions with pgbouncer and similar
    pool managers.
  - Fixed MCP per-user multi-header handling and OAuth token refresh.
  - Multiple Helm, OpenSearch, and tier-management fixes.
@@ -743,7 +769,7 @@

  ## v3.3.0

  Spreadsheets and CSVs become first-class searchable content, and a substantial round of observability, hardening,
  and polish. Thank you to our contributors for helping make this release possible!

  ## [Update] Defaults that Changed
@@ -759,9 +785,9 @@

  ### Tabular File Indexing

  Spreadsheets and CSVs are now indexed as structured tabular content rather than flattened text.
  A new section-based chunking pipeline emits field-value chunks, per-sheet descriptors, and whole-file summary chunks,
  so questions more accurately choose the right XLSX or CSV document, which the llm can then interpret.
  Tabular files are supported across the file upload path and the Google Drive, Blob, and SharePoint connectors.

  ### Smoother Chat Streaming
@@ -774,8 +800,8 @@

  - [@jaffarkeikei](https://github.com/jaffarkeikei) fixed inverted
    `already_existed` logic in the ingestion API response!
  - [@aserafin-mtt](https://github.com/aserafin-mtt) fixed MCP Langfuse
    tracing and Jira ADF parsing!
  - [@Mberga14](https://github.com/Mberga14) fixed a `Button`/`button`
    case-split that broke Linux builds and added support for extra manifests in the Helm chart!

@@ -792,20 +818,20 @@
  - Image extraction from PDF files.
  - Indexing is more resilient: staged-file handling, orphan file cleanup,
    and document processing that no longer dies mid-run on individual failures.
  - Option to run Onyx with OpenSearch only, fully disabling Vespa.

  ### Chat

  - Queued messages — send a follow-up while a response is still generating.
  - Multi-model chat refinements: inline citation chips in panels, isolated
    streaming errors per panel, and numerous UI polish fixes.
  - Scrollable tables with an overflow fade in chat responses.
  - GPT Image 2 support for image generation.
  - Anthropic Opus 4.7 added to the recommended models.

  ### Connectors

  - Gong connector converted from polling to checkpointed indexing.
  - HubSpot connector made significantly more efficient with fewer API calls.
  - Confluence: 504 retry with smaller page size, plus `/Date` macro and
    Cyrillic encoding fixes.
@@ -857,7 +883,7 @@

  The Hook system introduced in v3.1.0 has matured. A new admin page lets you create, edit,
  and monitor hooks with live status and logs,
  and a second integration point — document ingestion — joins query processing in the set of hookable lifecycle events.
  Connection-lost states surface clearly in the UI when a hook becomes unreachable.

  <img className="rounded-image" src="/assets/changelog/v3.2/hook_extensions_page.png" alt="Hook Extensions Page"/>
@@ -912,17 +938,17 @@

  ### Deployment and Observability

  - Grafana dashboard provisioning via Helm, with a dedicated OpenSearch
    retrieval dashboard
  - AWS deployment additions: RDS snapshots, EKS CloudWatch logging, RDS CPU
    / memory / IO alerts, and support for multiple Postgres hosts

  ### Notable Bug Fixes

  - Fixed owner-password-only PDFs being incorrectly rejected as protected
  - Fixed XLSX empty row/column handling
  - Fixed PowerPoint chart conversion failing during indexing
  - Fixed paused file connectors not reindexing correctly
  - Fixed MCP server propagating generic error messages instead of the real
    cause
  - Fixed mobile sidebar overlapping content on medium-sized screens, plus
@@ -1003,21 +1029,21 @@

  ### Connectors

  - SharePoint connector gained deduplication, filters/denylists, non-public
    shareable links, and better error logging
  - Notion connector upgraded to the latest API version
  - Confluence client retries on transient failures
  - Jira large-batch JSON decode fix
  - Google Drive: rate-limit retry and fallback access path for externally
    owned items
  - Discord connector now properly cleans up async resources

  ### Agents and Chat

  - Tool call arguments now stream live, with a delta-based jsonriver
    implementation
  - Custom tools have improved error handling and a refreshed timeline UI
  - Configurable file upload size limit, with a frontend precheck for oversized
    uploads. This can be configured in the Admin Panel, under Advanced Options for Chat Preferences.

    <img className="rounded-image" src="/assets/changelog/v3.1/file_attachment_size_limit_page.png" alt="File Attachment Size Limit"/>
@@ -1025,7 +1051,7 @@
  - More file types can be attached to chats, and the file preview supports
    significantly more formats.
    File types which are not explicitly supported/recognized are parsed using a text-detection heuristic.
  - Slackbot now resolves channel references and can filter search results by
    channel
  - Search state machine consolidated for more predictable behavior

@@ -1044,9 +1070,9 @@
  ### Performance

  - Stop deep-copying during indexing
  - Plaintext file results are now cached in chat
  - OpenSearch tuning continues with configurable shards/replicas, configurable
    Vespa migration page size, configurable hybrid subquery hits, vectors excluded from hybrid/random search payloads,
    and Prometheus metrics for retrieval

  ### Deployment
@@ -1075,39 +1101,39 @@

  ## v3.0.0

  The Vespa to OpenSearch migration is here, SCIM 2.0 lands for Enterprise,
  search and chat are unified into a single experience, Code Interpreter ships by default,
  and Onyx Lite and the Onyx CLI debut. v3.0.0 is the largest release we've ever cut.
  Thank you to our contributors for helping make this release possible!

  ## [Update] OpenSearch Migration

  v3.0.0 ships the full Vespa to OpenSearch migration toolchain.
  OpenSearch now runs alongside Vespa in the standard `docker-compose.yml` and Helm chart,
  and admins get a new Migration tab in the Admin Panel for kicking off and monitoring the migration with live progress.

  Vespa continues to serve all queries throughout the migration by default. Once the migration completes,
  admins can switch Onyx to read from OpenSearch.
  Setup details and guidance live in the [OpenSearch deployment
  guide](https://docs.onyx.app/deployment/local/opensearch),
  and the full migration walkthrough is in the [Vespa to OpenSearch migration
  guide](https://docs.onyx.app/admins/advanced_configs/opensearch_document_index_migration).

  ## Upgrade Path to v4.0.0

  v3.0.0 starts the final stretch of the Vespa to OpenSearch migration.
  **Every deployment must run a v3.x.x release long enough for the migration to complete before upgrading to v4.x.x,
  where Vespa will be fully removed.** If you skip v3 entirely,
  you will lose all of your indexed data and have to re-index from scratch — including re-uploading any Project files.

  ## Key Improvements

  ### SCIM 2.0 User and Group Provisioning

  Onyx Enterprise now ships a complete SCIM 2.0 implementation. Identity providers like Okta and Entra ID can create,
  update, and deactivate users and groups against Onyx via the standard `/scim/v2/*` endpoints with bearer-token auth.
  Admins can mint and rotate SCIM tokens from a new section in the Admin Panel,
  and we've passed Okta's Runscope spec test for OIN submission.
  [More details can be found in our documentation.](https://docs.onyx.app/deployment/authentication/scim#scim)

  ### Unified Search and Chat
@@ -1157,12 +1183,12 @@

  - Featured agents with labels can be highlighted in the agent picker without
    being public
  - New chat sharing UI with cleaner permissioning
  - Refreshed admin Chat Preferences page
  - Infinite scroll in the chat session sidebar
  - Configurable MCP host setting
  - Tab reading from the Chrome extension side panel
  - Tooltip and timeline polish, including better tool status messages and a fix
    for double-collapse in agent timelines
  - Throughout the codebase and UI, "assistant" identifiers have been renamed to
    "agent" for consistency, and "LLM Models" is now "Language Models"
@@ -1180,7 +1206,7 @@
  ### Deployment and Helm

  - Production Prometheus instrumentation for the API server, including
    SQLAlchemy connection pool and per-tenant metrics
  - Code Interpreter is included in the default Docker and Helm deployments
  - New Postgres-backed cache backend abstraction for caches that previously
    required Redis
@@ -1191,7 +1217,7 @@
    `toolConfig`
  - Fixed Vertex AI Claude Opus 4.6 model selection and sorted Vertex model
    names
  - Fixed Ollama context-window passthrough, content skipping, and DeepSeek
    handling
  - Fixed desktop app link clicking behavior
  - Fixed basic users being unable to share agents
@@ -1203,7 +1229,7 @@
  - Fixed search tool being enabled when no knowledge sources were selected
  - Fixed Google connectors redirecting to an auth error page after disconnect
  - Fixed Gong connector to respect the `Retry-After` header
  - Multiple UI fixes across modals, dropdowns, sidebar, whitelabel logo, code
    blocks, popovers, and the projects modal

  ---
@@ -1228,8 +1254,8 @@

  The v3.0.0 beta tag should be promoted very soon.

  As a reminder, everyone must upgrade to a v3.x.x version before moving on to v4.x.x,
  where Vespa will be fully removed.

  ## Key Improvements

@@ -1280,13 +1306,13 @@
  ### Notable Bug Fixes

  - Fixed Bedrock tool calling configuration
  - Fixed Ollama model selection in the UI
  - Fixed desktop app link clicking behavior
  - Fixed LLM provider configuration not syncing model flow data
  - Fixed search tool being enabled when no knowledge sources were selected
  - Fixed private models appearing in Agent creation
  - Fixed LLM read timeout handling
  - Fixed Asana connector workspace/team ID mismatch
  - Fixed Salesforce connector cleanup logic
  - Multiple UI fixes across modals, dropdowns, agent creation, and sidebar rendering

@@ -1308,11 +1334,11 @@

  ## [Update] OpenSearch Migration

  Our migration from Vespa to OpenSearch (announced in [v2.8.0 notes](#[announcement]-vector-database-migration))
  continues to make good progress.

  We've successfully migrated our internal Onyx instance to OpenSearch!
  The migration processed ~200k chunks/hour on a Helm-based cluster with 128GB Vespa (r6i.4xlarge nodes),
  8GB OpenSearch, and 8GB migration worker memory limits.

  Remaining work includes polishing the migration flow, improving performance, and final QA. We expect v3.0.0,
@@ -1370,11 +1396,11 @@
  - Fixed Azure OpenAI tool calling and LiteLLM Azure model streaming issues
  - Fixed streaming text appearing in bursts after citations
  - Fixed citation sidebar not appearing for web_search-only assistants
  - Fixed Ollama model selection in the UI
  - Fixed Asana connector workspace/team ID mismatch
  - Fixed Salesforce connector cleanup logic
  - Improved Slack query expansion to filter garbage from person names
  - Added Redis TTLs to prevent potential memory issues
  - Multiple UI fixes across modals, dropdowns, LLM provider configuration, and agent creation

  ---
@@ -1395,12 +1421,12 @@

  ## [Update] Vector Database Migration

  A small change in plans for our migration from Vespa to OpenSearch (announced in [v2.8.0
  notes](#[announcement]-vector-database-migration)).

  Instead of v2.11.0, the migration version will be **v3.0.0**. Vespa will be deprecated in **v4.0.0**.

  Our intention is that everyone must upgrade to a **v3.x.x** version before moving on to **v4.x.x**. Additionally,
  we do not anticipate many minor versions of v3 before moving to v4.

  <Note>
@@ -1409,9 +1435,9 @@
  </Note>

  Everything else remains the same. When you upgrade to a v3 version,
  admins will be prompted to start the migration process and Vespa will continue to power your search until v4.0.0.

  To accomodate this change, our timeline has been pushed by 2 weeks:
  - **v3.0.0 Beta** will be released in the second week of February
  - **v3.0.0 Stable** is slated for the third week of February

@@ -1433,12 +1459,12 @@

  <img className="rounded-image" src="/assets/admins/discord_bot/register_server.png" alt="Onyx Discord Bot"/>

  ### Nano Banana Image Generation

  You can now configure Nano Banana as an image generation provider.

  Add Vertex AI credentials in the **Image Generation** page of the Admin Panel to use Gemini 2.5 Flash Image (Nano
  Banana) and Gemini 3 Pro Image (Nano Banana Pro).

  <img className="rounded-image" src="/assets/changelog/v2.10/showcase_nano_banana.png" alt="Nano Banana Image Generation"/>

@@ -1461,7 +1487,7 @@

  <Info>
    Note that if you switch from the federated connector,
    you will lose access to searching group DMs and direct messages.
  </Info>

  Read more about the [Slack connector](/admins/connectors/official/slack/slack_indexed)
@@ -1512,7 +1538,7 @@
  Features include:
  - **Quick-access address bar** - Type your question directly in the address bar
  - **New-tab page** - Start every browsing session with Onyx
  - **Contextual insights** - Highlight any text on a webpage and birng it into your chat
  - **Always-on sidebar** - Open Onyx as a sidebar on any webpage

  Check out the [Chrome Extension](/overview/onyx_anywhere/chrome_extension) page to get started!
@@ -1521,7 +1547,7 @@

  ### Onyx Widget

  You can now embed an Onyx chatbot into any website!
  Easily add an AI agent or knowledge search to your website with a single line of code.

  The widget abstracts our chat API and can be configured to use our Cloud or your self-hosted instance of Onyx.
@@ -1538,7 +1564,7 @@

  ## Contributor Highlights

  - [@coreyauger](https://github.com/coreyauger) updated some out-of-date links in our READMEs!

  ## Additional Improvements

@@ -1568,18 +1594,18 @@

  ## [Announcement] Vector Database Migration

  We've recently made the decision to migrate the Onyx vector database from Vespa to OpenSearch.
  OpenSearch is a FOSS fork of ElasticSearch, and will be deployed similarly to the existing `index` container.
  We expect OpenSearch to be easier to develop with, provide equal if not better search performance, and, importantly,
  have a smaller memory footprint.

  <Note>
    The official migration from Vespa to OpenSearch will begin in February, 2026, with the **v2.11.0** release.
  </Note>

  In v2.11.0, admins will be prompted to start a migration process,
  which will copy over all indexed data from Vespa to OpenSearch.
  Onyx will remain fully functional with Vespa in this version.

  Following v2.11.0, we plan to release v3.0.0, which will use OpenSearch exclusively.

@@ -1604,7 +1630,7 @@

  ### Cost Tracking

  We've added LLM cost tracking and time to first token metrics for tracing providers like Braintrust.

  See our [deployment configuration docs](/deployment/configuration/configuration#observability)
  for how to set up observability for your deployment.
@@ -1660,7 +1686,7 @@
  There's a new page for configuring image generation models in the Admin Panel.

  You can re-use your existing OpenAI API key or setup new OpenAI or Azure providers.
  Nano Banana support is coming soon!

  <Info>
    If you've previously set up an image generation model via environment variables,
@@ -1690,7 +1716,7 @@
  You will no longer see the container start and stop (when disabled) in Docker.

  To deploy the MCP Server,
  you will need to set `MCP_SERVER_ENABLED=true` in your environment and uncomment the MCP Server section in your
  `docker-compose.yml` file.

  ### Model Server Changes
@@ -1825,9 +1851,9 @@

  <img className="rounded-image" src="/assets/changelog/v2.6/showcase_improved_file_handling.png" alt="Improved File Handling in Chat"/>

  ### MCP Passthrough Authentication

  MCP connections now support passthrough authentication for seamless integration with your existing auth systems.

  ### Embedding Model Configuration

@@ -1835,7 +1861,7 @@

  ### Bug Fixes

  - Multiple fixes to Confluence, Jira, and web connectors
  - Improved document processing stability, reducing likelihood of OOM errors when using API-based embedding models

  ---
@@ -1858,7 +1884,7 @@

  ### Web Search Enhancements

  We've officially released support for new web search providers including Google PSE, Serper, Firecrawl,
  and a built-in Onyx web scraper!

  Configure these providers via the Web Search page in the Admin Panel and give your agents the ability to search the
@@ -1870,7 +1896,7 @@

  ### Claude Opus 4.5 Support

  Onyx now supports Anthropic's newest, most powerful model: Claude Opus 4.5!
  Configure Claude Opus 4.5 via the LLM Configuration page in the Admin Panel.

  <img className="rounded-image" src="/assets/changelog/v2.5/showcase_45_opus.png" alt="Claude Opus 4.5"/>
@@ -1899,8 +1925,8 @@

  ## Contributor Highlights

  - [@sashank-rayapudi-ai](https://github.com/sashank-rayapudi-ai) added a new connector for Testrail!
  - [@mristau-alltrails](https://github.com/mristau-alltrails) identified and resolved a vulnerability in our nginx configuration!

  ## Additional Improvements

@@ -1919,9 +1945,9 @@
  Connectors that continuously fail will now automatically be paused. This helps relieve pressure on the background,
  API, and indexing containers, improving overall system stability.

  ### Kubernetes Affinity and Tolerations

  We've added affinity and tolerations to the Helm deployment templates,
  allowing you to customize where pods are scheduled in your Kubernetes environment.

  ---
@@ -1950,12 +1976,12 @@

  Configure filters such as:
  - Channel selection with glob patterns
  - Message type filtering (1:1 DMs, group DMs, private channels)
  - Date extraction

  <Note>
    If you have an existing Slack Federated Connector,
    you will need to update the app's scopes to use the expanded search functionality (e.g. searching group DMs).

    Additionally,
    each user will need to revoke and re-authorize their OAuth access to the app for the new scopes to take effect.
@@ -2067,13 +2093,13 @@

  ### More Resilient Gmail Connector

  The Gmail connector now uses checkpointing for more resilient indexing.
  This makes the connector more reliable and significantly cheaper to run,
  especially when re-embedding your entire Gmail history.

  ### Air-Gapped Deployment Support

  We've fixed air-gapped deployments by packaging the tokenizer directly with the model server image.
  This ensures Onyx can run in completely isolated environments without requiring external dependencies.

  ### Optimized LLM Interactions
@@ -2114,7 +2140,7 @@
  ### New Login UI and Onboarding Flow

  First-time users now experience a streamlined onboarding journey that guides them through personalizing their
  workspace, setting up their LLMs, and configuring actions like Web Search and Image Generation.

  <img className="rounded-image" src="/assets/changelog/v2.2/showcase_onboarding.png" alt="Onboarding flow"/>

@@ -2144,14 +2170,14 @@

  ## Additional Improvements

  ### Confluence OAuth and Jira Fixes

  We've resolved issues with Confluence OAuth authentication and updated Jira integration to address deprecated API
  endpoints, ensuring reliable connectivity with your Atlassian tools.

  ### Enhanced Provider Support

  Onyx now has improved support for LLM providers like Ollama and OpenRouter,
  giving you more flexibility in choosing the models that work best for your use case.

  ---

diff --git a/docs.json b/docs.json
@@ -34,6 +34,7 @@
               "overview/core_features/code_interpreter",
               "overview/core_features/image_generation",
               "overview/core_features/voice_mode",
+              "overview/core_features/general_settings",
               {
                 "group": "Craft",
                 "icon": "hammer",

diff --git a/overview/core_features/general_settings.mdx b/overview/core_features/general_settings.mdx
@@ -0,0 +1,32 @@
+---
+title: "General Settings"
+description: "Profile and interface preferences"
+icon: "gear"
+---
+
+## Overview
+
+The **General** tab of your User Settings page holds account-level preferences that apply across your Onyx experience,
+including your interface language.
+
+<img className="rounded-image" src="/assets/overview/GENERIC_PLACEHOLDER.png" alt="Onyx General Settings"/>
+
+## Interface Language
+
+Onyx's interface is available in:
+
+- English (default)
+- Español
+- Português
+- Français
+- Deutsch
+
+<Note>
+  Language names are always shown in their native spelling, regardless of which language is currently active,
+  so you can find your language even if you can't read the current interface text.
+</Note>
+
+To change your interface language, open **User Settings > General** (`/app/settings/general`)
+and select a language from the **Language** dropdown.
+
+<img className="rounded-image" src="/assets/overview/GENERIC_PLACEHOLDER.png" alt="Onyx Language dropdown"/>