# Jan changelog > Run large language models locally, offline. - Vendor: Menlo Research - Category: AI - Platforms: Desktop - Official site: https://jan.ai - Tracked by: What's New (https://whatsnew.fyi/product/jan) - Harvested from: GitHub (menloresearch/jan) - Entries below: 10 (newest first) What's New is an index, not a publisher: every entry below links to the vendor's own release notes, which are the authoritative source. Entries are labelled where they are hand-curated sample data, pre-releases, or drawn from a secondary source such as a developer blog. ## Releases ### v0.8.4 — 0.8.4 - Date: 2026-07-23 - Version: v0.8.4 - Original notes: https://github.com/janhq/jan/releases/tag/v0.8.4 - Permalink: https://whatsnew.fyi/product/jan/releases/v0.8.4 - Labels: Platforms: Desktop - **changed** — Settings and credentials now live in a backend-managed store instead of localStorage, with automatic migration of existing settings and API keys on first launch - **added** — Toggle for folding interim text into reasoning trace in chat - **fixed** — Repair embeddings, stop chat-model reloads, and parse more PDFs in RAG - **fixed** — Stop persisting detached assistant replies in branched threads - **changed** — Move settings and credentials off localStorage to backend store with secrets saved to OS keyring - **fixed** — Coerce numeric assistant params from string to number - **fixed** — Use snake case for ADL enum in get_gpu_usage - **fixed** — Stop MCP tool-approval popup from hanging and move transient approval state off disk - **added** — API key rotation for Gemini - **security** — Bump handlebars to 4.7.9 to clear RCE advisory - **fixed** — Ignore custom SSE event types in OpenAI-compatible streams - **added** — OpenAI-compatible translating gateway, Responses API, and unified reasoning - **fixed** — Stabilize chat prefix caching, restore cloud providers, and unify logging - **fixed** — Guard against duplicate serve() on streamable-http MCP servers - **fixed** — Send anthropic-version on model-list fetch for Anthropic provider - **changed** — macOS hide-on-close behavior and hide-to-tray when Local API Server runs - **added** — Bundle built-in extensions into the app binary - **added** — Per-model chat template kwargs for llamacpp - **added** — Token counter for remote and MLX providers - **added** — Native web_search and web_fetch via tauri-plugin-websearch - **fixed** — Preserve AbortError identity when stopping mid model-load in chat - **fixed** — Stop hiding the llama.cpp token counter during chat - **fixed** — Persist chain-of-thought duration in thread metadata - **fixed** — Force code blocks LTR in RTL language context for markdown ##### Migration **Settings and credentials now live in a backend-managed store.** Provider settings and credentials previously persisted in the webview's `localStorage`. In 0.8.4, Jan reads and writes them from a backend-managed store instead (secrets go to the OS keyring). Your existing settings and API keys are copied over automatically on first launch - no manual action is needed, and from this version on all new changes are saved to the backend store. Your old `localStorage` data is **not deleted** - it's kept as a snapshot so you can downgrade to a pre-0.8.4 build and still have your previous settings. Any changes you make in 0.8.4+ live only in the backend store, so a downgrade sees the older snapshot, not your latest settings. --- ##### What's Changed * chore(flatpak): add 0.8.3 release notes to appdata by @qnixsynapse in https://github.com/janhq/jan/pull/8343 * docs: update changelogs for v0.8.3 and update documentation by @qnixsynapse in https://github.com/janhq/jan/pull/8336 * Merge release/v0.8.3 to main by @qnixsynapse in https://github.com/janhq/jan/pull/8344 * i18n(ja): update Japanese translations (#8264) by @mahirhir in https://github.com/janhq/jan/pull/8348 * i18n(ja): complete common.json translations and fix terminology (#8264) by @mahirhir in https://github.com/janhq/jan/pull/8349 * i18n(ja): complete settings.json translations (#8264) by @mahirhir in https://github.com/janhq/jan/pull/8352 * fix(markdown): force code blocks LTR in RTL language context by @qnixsynapse in https://github.com/janhq/jan/pull/8350 * feat(chat): add toggle for folding interim text into reasoning trace by @qnixsynapse in https://github.com/janhq/jan/pull/8363 * fix(rag): repair embeddings, stop chat-model reloads, and parse more PDFs by @qnixsynapse in https://github.com/janhq/jan/pull/8360 * ci: close issues inactive for over 180 days by @qnixsynapse in https://github.com/janhq/jan/pull/8366 * ci: grant actions:write to stale workflow for state persistence by @qnixsynapse in https://github.com/janhq/jan/pull/8377 * fix(chat): stop persisting detached assistant replies in branched threads by @qnixsynapse in https://github.com/janhq/jan/pull/8376 * feat(settings): move settings/credentials off localStorage to backend store by @qnixsynapse in https://github.com/janhq/jan/pull/8388 * fix(params): coerce numeric assistant params from string to number by @qnixsynapse in https://github.com/janhq/jan/pull/8386 * fix: use snake case for ADL enum in get_gpu_usage by @nkulshre-amd in https://github.com/janhq/jan/pull/8390 * fix(chat): stop MCP tool-approval popup from hanging; move transient approval state off disk by @qnixsynapse in https://github.com/janhq/jan/pull/8391 * API key rotation for Gemini by @jjchen01 in https://github.com/janhq/jan/pull/8408 * fix(security): bump handlebars to 4.7.9 to clear RCE advisory by @qnixsynapse in https://github.com/janhq/jan/pull/8395 * fix(chat): ignore custom SSE event types in OpenAI-compatible streams by @qnixsynapse in https://github.com/janhq/jan/pull/8392 * feat: OpenAI-compatible translating gateway + Responses API + unified reasoning by @qnixsynapse in https://github.com/janhq/jan/pull/8423 * fix: Stabilize chat prefix caching, restore cloud providers, unify logging (v0.8.4) by @qnixsynapse in https://github.com/janhq/jan/pull/8425 * fix(mcp): guard against duplicate serve() on streamable-http MCP servers by @thinhlpg in https://github.com/janhq/jan/pull/8435 * fix(providers): send anthropic-version on model-list fetch by @qnixsynapse in https://github.com/janhq/jan/pull/8433 * fix(desktop): macOS hide-on-close; hide-to-tray when Local API Server runs by @qnixsynapse in https://github.com/janhq/jan/pull/8436 * feat(extensions): bundle built-in extensions into the app binary and update tests by @qnixsynapse in https://github.com/janhq/jan/pull/8437 * feat(llamacpp): per-model chat template kwargs by @qnixsynapse in https://github.com/janhq/jan/pull/8441 * feat: token counter for remote/ML _[Truncated at 4000 characters — full notes: https://github.com/janhq/jan/releases/tag/v0.8.4]_ ### v0.8.3 — 0.8.3 - Date: 2026-06-24 - Version: v0.8.3 - Original notes: https://github.com/janhq/jan/releases/tag/v0.8.3 - Permalink: https://whatsnew.fyi/product/jan/releases/v0.8.3 - Labels: Platforms: Desktop - **fixed** — Use Windows copy command in build-cli Makefile target - **fixed** — Browser port leak and auto-disable UX in MCP - **fixed** — Correct sampler defaults, fix sheet layout under window controls, and harden backend installs - **fixed** — Allow data:/blob: media in CSP so video uploads load - **fixed** — Render lone SVG in a code fence; do not extract SVG nested in real code - **fixed** — Render media on first-message send in chat ##### What's Changed * docs: add v0.8.2 changelog by @qnixsynapse in https://github.com/janhq/jan/pull/8246 * merge release/v0.8.2 to main by @qnixsynapse in https://github.com/janhq/jan/pull/8250 * chore: add flatpak manifest 0.8.2 by @Minh141120 in https://github.com/janhq/jan/pull/8259 * fix: use Windows copy command in build-cli Makefile target by @ianmohney in https://github.com/janhq/jan/pull/8293 * Updated translation by @albanobattistella in https://github.com/janhq/jan/pull/8262 * fix(mcp): browser port leak and auto-disable ux by @abn in https://github.com/janhq/jan/pull/8317 * Merge dev/v0.8.3 to main by @qnixsynapse in https://github.com/janhq/jan/pull/8310 * docs: clarify Jan model quick-start flow by @archit-goyal in https://github.com/janhq/jan/pull/8323 * Fix: correct sampler defaults, fix sheet layout under window controls, and harden backend installs by @qnixsynapse in https://github.com/janhq/jan/pull/8327 * fix(csp): allow data:/blob: media so video uploads load by @qnixsynapse in https://github.com/janhq/jan/pull/8330 * fix(markdown): render lone in a code fence; don't extract nested in real code by @qnixsynapse in https://github.com/janhq/jan/pull/8332 * fix(chat): render media on first-message send by @qnixsynapse in https://github.com/janhq/jan/pull/8337 ##### New Contributors * @ianmohney made their first contribution in https://github.com/janhq/jan/pull/8293 * @abn made their first contribution in https://github.com/janhq/jan/pull/8317 * @archit-goyal made their first contribution in https://github.com/janhq/jan/pull/8323 **Full Changelog**: https://github.com/janhq/jan/compare/v0.8.2...v0.8.3 ### v0.8.2 — 0.8.2 - Date: 2026-06-01 - Version: v0.8.2 - Original notes: https://github.com/janhq/jan/releases/tag/v0.8.2 - Permalink: https://whatsnew.fyi/product/jan/releases/v0.8.2 - Labels: Platforms: Desktop - **fixed** — Honor configured attachment size limit in document parser - **fixed** — Register mlx capability so its commands pass ACL on macOS - **fixed** — Persist assistant parameters so sampler settings survive restart - **fixed** — Stop Tauri nonce-ing style-src so unsafe-inline works - **fixed** — Raise default ctx-size to 8192 for llamacpp - **fixed** — Move toast accent tint to CSS class to satisfy CSP - **fixed** — Restore toast styling on Windows and tint toast with accent color - **fixed** — Make notification toast visible and theme-aware - **fixed** — Add Italian to language switcher and fix Catalan locale code - **fixed** — Surface friendly message on transport-level request failures in providers - **added** — Add copy and reveal buttons to API key inputs in providers - **fixed** — Bind requestIdleCallback to window to avoid Illegal invocation in updater - **changed** — Defer backend dependency verification to browser idle for llamacpp - **changed** — Defer backend update check to browser idle in updater - **changed** — Defer initial app update check until after first paint in updater - **fixed** — Start router before deferring backend update check in llamacpp - **fixed** — List MLX models in Local API Server default model picker - **fixed** — Make attachment chip text legible in user bubble - **added** — Add toggle for automatic app update check in settings - **fixed** — Make the thread's agent authoritative so in-chat switching works ##### What's Changed - fix(rag): honor configured attachment size limit in document parser by @qnixsynapse in https://github.com/janhq/jan/pull/8248 - fix(mlx): register mlx capability so its commands pass ACL on macOS by @qnixsynapse - fix(threads): persist assistant parameters so sampler settings survive restart by @qnixsynapse - fix(csp): stop Tauri nonce-ing style-src so unsafe-inline works by @qnixsynapse - test(extensions): run llamacpp-extension tests in the global suite by @qnixsynapse - fix(llamacpp): raise default ctx-size to 8192 by @qnixsynapse - fix(ui): move toast accent tint to CSS class to satisfy CSP by @qnixsynapse - fix(ui): restore toast styling on Windows; tint toast with accent color by @qnixsynapse - fix(ui): make notification toast visible and theme-aware by @qnixsynapse - fix(i18n): add Italian to language switcher, fix Catalan locale code by @qnixsynapse - fix(providers): surface friendly message on transport-level request failures by @qnixsynapse - feat(providers): add copy and reveal buttons to API key inputs by @qnixsynapse - fix(updater): bind requestIdleCallback to window to avoid Illegal invocation by @qnixsynapse - perf(llamacpp): defer backend dep verification to browser idle by @qnixsynapse - perf(updater): defer backend update check to browser idle by @qnixsynapse - build(mlx): build the MLX plugin and extension on macOS only by @qnixsynapse - perf(updater): defer initial app update check until after first paint by @qnixsynapse - fix(llamacpp): start router before deferring backend update check by @qnixsynapse - fix(server): list MLX models in Local API Server default model picker by @qnixsynapse - fix(chat): make attachment chip text legible in user bubble by @qnixsynapse - feat(settings): add toggle for automatic app update check by @qnixsynapse - fix(assistant): make the thread's agent authoritative so in-chat switching works by @qnixsynapse - fix(startup): increase logo–caption gap and center loader caption by @qnixsynapse - feat(downloads): pause and resume model downloads from the hub panel by @qnixsynapse - fix(chat): route prompt-too-big errors to the Increase Context Size banner by @qnixsynapse - fix(llamacpp): recover crashed models and surface mid-request crashes by @qnixsynapse - feat(startup): branded loader with phased status; defer MCP past first paint by @qnixsynapse - fix(linux): keep tao's draggable Wayland titlebar; follow GNOME button-layout by @qnixsynapse - feat(llamacpp): add HIP (AMD ROCm) backend support by @qnixsynapse - fix(llamacpp): unblock UI from router startup; watchdog against blank window by @qnixsynapse - style(message): use primary color for user message bubble by @qnixsynapse - chore(release): bump app and plugin versions to 0.8.2 by @qnixsynapse **Full Changelog**: https://github.com/janhq/jan/compare/v0.8.1...v0.8.2 ### v0.8.1 — 0.8.1 - Date: 2026-05-29 - Version: v0.8.1 - Original notes: https://github.com/janhq/jan/releases/tag/v0.8.1 - Permalink: https://whatsnew.fyi/product/jan/releases/v0.8.1 - Labels: Platforms: Desktop - **added** — Italian translation - **fixed** — Clear deletedModels tombstones for re-downloaded models - **fixed** — Use OS-native TLS trust store on desktop - **fixed** — Disable auto-fit by default and set default ctx_len to 8192 for llamacpp - **fixed** — Skip NVIDIA/Vulkan GPU probes on macOS - **fixed** — Avoid router invariant in PromptProgress for threads - **fixed** — Restore Attachments entry in settings menu - **fixed** — Guard null entity arrays in TweetCard ##### What's Changed * Add Italian Translation by @albanobattistella in https://github.com/janhq/jan/pull/8157 * chore: merge release v0.8.0 to main by @qnixsynapse in https://github.com/janhq/jan/pull/8170 * chore: fix build due to missing comma in italian i18n by @qnixsynapse in https://github.com/janhq/jan/pull/8171 * docs: update for v0.8.0 by @qnixsynapse in https://github.com/janhq/jan/pull/8164 * chore: update manifest and changelog release 0.8.0 by @Minh141120 in https://github.com/janhq/jan/pull/8172 * docs: remove connectors note from 0.8.0 changelog by @qnixsynapse in https://github.com/janhq/jan/pull/8173 * chore: update filename changelog by @Minh141120 in https://github.com/janhq/jan/pull/8177 * fix(docs): guard null entity arrays in TweetCard by @qnixsynapse in https://github.com/janhq/jan/pull/8187 * fix: clear deletedModels tombstones for re-downloaded models by @qnixsynapse in https://github.com/janhq/jan/pull/8178 * chore(ci): pin all GitHub Actions to SHAs and add cancel-in-progress by @Minh141120 in https://github.com/janhq/jan/pull/8193 * fix(tls): use OS-native TLS trust store on desktop by @FyrbyAdditive in https://github.com/janhq/jan/pull/8186 * chore(ci): bump dcarbone/install-jq-action from v2.0.1 to v3.2.0 by @Minh141120 in https://github.com/janhq/jan/pull/8199 * chore: merge dev branch fix/macos backend-install-ux by @qnixsynapse in https://github.com/janhq/jan/pull/8197 * fix(llamacpp): disable auto-fit by default, default ctx_len to 8192 by @qnixsynapse in https://github.com/janhq/jan/pull/8204 * Merge fix/pre-r-dev branch to main by @qnixsynapse in https://github.com/janhq/jan/pull/8207 * Fix/system theme launch by @qnixsynapse in https://github.com/janhq/jan/pull/8217 * fix(hardware): skip NVIDIA/Vulkan GPU probes on macOS by @qnixsynapse in https://github.com/janhq/jan/pull/8219 * fix(threads): avoid router invariant in PromptProgress by @qnixsynapse in https://github.com/janhq/jan/pull/8220 * fix(settings): restore Attachments entry in settings menu by @qnixsynapse in https://github.com/janhq/jan/pull/8222 ##### New Contributors * @albanobattistella made their first contribution in https://github.com/janhq/jan/pull/8157 * @FyrbyAdditive made their first contribution in https://github.com/janhq/jan/pull/8186 **Full Changelog**: https://github.com/janhq/jan/compare/v0.8.0...v0.8.1 ### v0.8.0 — 0.8.0 - Date: 2026-05-22 - Version: v0.8.0 - Original notes: https://github.com/janhq/jan/releases/tag/v0.8.0 - Permalink: https://whatsnew.fyi/product/jan/releases/v0.8.0 - Labels: Platforms: Desktop - **added** — Add MiniMax as a predefined cloud provider with M2.7 default - **changed** — Replace Foundation Models Swift HTTP server with direct Rust FFI via fm-rs - **fixed** — Add pointer cursor to tool header when hovering - **fixed** — Make Makefile compilable from powershell Windows - **fixed** — Resolve MCP stdio transport disconnect and add auto-reconnect - **added** — Add assistant chat button and fix random assistant issue - **fixed** — Latest model toast check with better logic - **added** — Add /v1/orchestrations for server-side MCP tool execution - **added** — Add GUI toggle for server-side tool execution on /chat/completi - **fixed** — Auto-approve internal RAG tools for embedded documents - **fixed** — Prevent data loss on app reload - **fixed** — Fix OpenAI chat crash in web runtime (invoke undefined) - **added** — Add paragraph-level Edit with AI for assistant markdown - **added** — Add Korean localization - **added** — Add translation to Catalan language - **fixed** — Fix missing redirects - **added** — Orchestrate MCP tools - route servers, avoid loading all tools - **added** — Auto-summarize thread titles after first response with cancellable background inference - **fixed** — Align KV cache default settings with actual runtime behavior - **added** — Add configurable token limits and auto-compact for remote model providers ##### What's Changed * feat: add MiniMax as a predefined cloud provider with M2.7 default by @octo-patch in https://github.com/janhq/jan/pull/7738 * feat: replace Foundation Models Swift HTTP server with direct Rust FFI via fm-rs by @dev-miro26 in https://github.com/janhq/jan/pull/7781 * fix: add pointer cursor to tool header when hovering by @bytrangle in https://github.com/janhq/jan/pull/7785 * docs: changelog for v0.7.9 by @Vanalite in https://github.com/janhq/jan/pull/7795 * chore: resolve conflict of main & release v0.7.9 by @Vanalite in https://github.com/janhq/jan/pull/7797 * chore: merge release v0.7.9 to main by @Vanalite in https://github.com/janhq/jan/pull/7798 * fix: Making Makefile compilable from powershell Windows by @thewulf7 in https://github.com/janhq/jan/pull/7771 * Chore/update template windows ci by @Minh141120 in https://github.com/janhq/jan/pull/7805 * fix: resolve MCP stdio transport disconnect and add auto-reconnect by @dev-miro26 in https://github.com/janhq/jan/pull/7791 * feat: add assistant chat button and fix random assistant issue by @markmehere in https://github.com/janhq/jan/pull/7750 * fix: latest model toast check with better logic by @Vanalite in https://github.com/janhq/jan/pull/7806 * feat: add /v1/orchestrations for server-side MCP tool execution by @claytonlin1110 in https://github.com/janhq/jan/pull/7800 * feat: add GUI toggle for server-side tool execution on /chat/completi… by @claytonlin1110 in https://github.com/janhq/jan/pull/7808 * fix: auto-approve internal RAG tools for embedded documents by @claytonlin1110 in https://github.com/janhq/jan/pull/7809 * docs: require screenshot or video record by @Vanalite in https://github.com/janhq/jan/pull/7811 * fix: prevent data loss on app reload by @linuxid10t in https://github.com/janhq/jan/pull/7810 * Fix OpenAI chat crash in web runtime (`invoke` undefined) by @dataCenter430 in https://github.com/janhq/jan/pull/7815 * feat: paragraph-level Edit with AI for assistant markdown by @claytonlin1110 in https://github.com/janhq/jan/pull/7812 * Update README to fix build commands by @SuperCowProducts in https://github.com/janhq/jan/pull/7829 * Add Korean localization by @hooray804 in https://github.com/janhq/jan/pull/7767 * Add translation to Catalan language by @anebot in https://github.com/janhq/jan/pull/7319 * fix: missing redirects by @louis-jan in https://github.com/janhq/jan/pull/7837 * feat: orchestrate MCP tools - route servers, avoid loading all tools by @dataCenter430 in https://github.com/janhq/jan/pull/7793 * feat: auto-summarize thread titles after first response with cancellable background inference by @statxc in https://github.com/janhq/jan/pull/7816 * fix: align KV cache default settings with actual runtime behavior by @linuxid10t in https://github.com/janhq/jan/pull/7832 * feat: add configurable token limits and auto-compact for remote model providers by @dev-miro26 in https://github.com/janhq/jan/pull/7799 * feat: enhance Tauri opener service for Linux file exploration by @Dexterity104 in https://github.com/janhq/jan/pull/7827 * feat(gguf): fix model compatibility heuristic for Apple Silicon unified memory by @dataCenter430 in https://github.com/janhq/jan/pull/7842 * Update Settings sidebar grouping and rename MCP to Connectors by @claytonlin1110 in https://github.com/janhq/jan/pull/7839 * feat/fix: add regenerate button on failed message responses by @linuxid10t in https://github.com/janhq/jan/pull/7844 * Updated Arm64 linux build instructions for latest app release + GPU acceleration by @SuperCowProducts in https://github.com/janhq/jan/pull/7841 * fix: let users scroll and read thinking content during streaming without being interrupted by @bittoby in https://github.com/janhq/jan/pull/7843 * chore: update icon privacy and code claude by @urmauur in https://github.com/janhq/jan/pull/7853 * fix: improve model selection bar usability by @claytonlin1110 in https://github.com/janhq/jan/pull/7852 * _[Truncated at 4000 characters — full notes: https://github.com/janhq/jan/releases/tag/v0.8.0]_ ### v0.7.9 — 0.7.9 - Date: 2026-03-23 - Version: v0.7.9 - Original notes: https://github.com/janhq/jan/releases/tag/v0.7.9 - Permalink: https://whatsnew.fyi/product/jan/releases/v0.7.9 - Labels: Platforms: Desktop - **fixed** — Clean up ANTHROPIC_* environment variables from shell config when server stops or app exits - **fixed** — Resolve Anthropic 400 error on serial tool calls - **fixed** — System theme not updating when switching dark/light mode on KDE Linux - **fixed** — GPU detection losing - **fixed** — Prevent duplicate backend downloads during llamacpp update - **fixed** — Render code block without language specification correctly - **changed** — Refactor normalizeLatex logic and improve performance - **fixed** — Update CLI help examples to use Jan model ID - **fixed** — KV-cache defaults and fit migration - **added** — Default model for local API server - **removed** — Remove the 32k context length enforcement upon API server start - **added** — Add default assistant setting - **added** — Add foundation models plugin with server management capabilities - **added** — Add Spanish (Español) locale - **fixed** — Change data location that has accidental symlink failures - **fixed** — Resolve downloaded model toggle bug by handling developer prefix - **added** — Safe define remove directory list during the factory reset - **removed** — Remove openclaw installation - **fixed** — CLI install method on Windows - **added** — Add NVIDIA NIM as a predefined cloud provider - **security** — Fix API server CORS vulnerability - **changed** — Replace toggle with chevron for show variants in Hub - **added** — Settings to auto increase context length - **added** — Expand supported file types for document attachments - **fixed** — Hub behavior issues - **added** — Remote model update - **fixed** — Auto-increase context size bugs - **fixed** — Restore default localhost on API Server start ##### What's Changed * fix: clean up ANTHROPIC_* env vars from shell config when server stops or app exits by @since-2017-hub in https://github.com/janhq/jan/pull/7628 * fix: add redirect from /guides to /docs by @louis-jan in https://github.com/janhq/jan/pull/7669 * docs: update troubleshooting guides by @louis-jan in https://github.com/janhq/jan/pull/7670 * docs: restructure Reference by @louis-jan in https://github.com/janhq/jan/pull/7674 * docs: agents availability by @louis-jan in https://github.com/janhq/jan/pull/7682 * docs: add community cards to desktop overview by @louis-jan in https://github.com/janhq/jan/pull/7687 * chore: resolve release main conflict 0.7.8 by @Vanalite in https://github.com/janhq/jan/pull/7713 * Sync release/v0.7.8 to main branch by @louis-jan in https://github.com/janhq/jan/pull/7711 * doc: add changelog for v0.7.8 by @Vanalite in https://github.com/janhq/jan/pull/7714 * docs: update CLI and agents docs for v0.7.8 by @louis-jan in https://github.com/janhq/jan/pull/7715 * feat: update flatpak manifest v0.7.8 by @Minh141120 in https://github.com/janhq/jan/pull/7717 * fix: resolve Anthropic 400 error on serial tool calls by @atoz96 in https://github.com/janhq/jan/pull/7696 * Revert: clear ANTHROPIC_ env on app exit #7628 by @louis-jan in https://github.com/janhq/jan/pull/7718 * Merge pull request #7695 from oolokioo7/fix/edit-dialog-width by @louis-jan in https://github.com/janhq/jan/pull/7719 * Fix: gpu detection losing by @dataCenter430 in https://github.com/janhq/jan/pull/7605 * fix: system theme not updating when switching dark/light mode on KDE Linux by @since-2017-hub in https://github.com/janhq/jan/pull/7629 * docs: add Windows data folder relocation troubleshooting section by @louis-jan in https://github.com/janhq/jan/pull/7724 * docs: add troubleshooting for timeout port by @Vanalite in https://github.com/janhq/jan/pull/7730 * fix: fix error on openclaw doc by @Vanalite in https://github.com/janhq/jan/pull/7731 * fix: prevent duplicate backend downloads during llamacpp update by @atoz96 in https://github.com/janhq/jan/pull/7721 * fix: render code block without lang spec correcty by @cs-cat in https://github.com/janhq/jan/pull/7741 * refactor: fix `normalizeLatex` logic and improve performance by @cs-cat in https://github.com/janhq/jan/pull/7743 * docs: troubleshooting update for jan CLI on Windows by @louis-jan in https://github.com/janhq/jan/pull/7736 * docs: add instructions to uninstall openclaw by @Vanalite in https://github.com/janhq/jan/pull/7754 * docs: Add instruction to install HIP/ROCm llamacpp backend for AMD by @Vanalite in https://github.com/janhq/jan/pull/7755 * fix: update CLI help examples to use Jan model ID by @louis-jan in https://github.com/janhq/jan/pull/7753 * fix: kv-cache defaults and fit migration by @louis-jan in https://github.com/janhq/jan/pull/7751 * feat: default model for local api server by @urmauur in https://github.com/janhq/jan/pull/7735 * chore: remove the 32k ctx_len enforcement upon API server start by @Vanalite in https://github.com/janhq/jan/pull/7756 * feat: add default assistant setting by @urmauur in https://github.com/janhq/jan/pull/7725 * doc: update CONTRIBUTING.md by @Vanalite in https://github.com/janhq/jan/pull/7764 * fix: wrong display shortcut create project by @urmauur in https://github.com/janhq/jan/pull/7757 * feat: add foundation models plugin with server management capabilities by @dev-miro26 in https://github.com/janhq/jan/pull/7744 * feat: add Spanish (Español) locale by @dev-miro26 in https://github.com/janhq/jan/pull/7765 * fix: change data location that has accident symlink failed by @Vanalite in https://github.com/janhq/jan/pull/7763 * fix: resolve downloaded model toggle bug by handling developer prefix by @jamesdam in https://github.com/janhq/jan/pull/7648 * feat: safe define remove dir list during the factory reset by @Vanalite in https://github.com/janhq/jan/pull/7766 * chore: remove ope _[Truncated at 4000 characters — full notes: https://github.com/janhq/jan/releases/tag/v0.7.9]_ ### v0.7.8 — 0.7.8 - Date: 2026-03-11 - Version: v0.7.8 - Original notes: https://github.com/janhq/jan/releases/tag/v0.7.8 - Permalink: https://whatsnew.fyi/product/jan/releases/v0.7.8 - Labels: Platforms: Desktop - **added** — Add xAI provider - **added** — Add OpenClaw integration - **added** — Add Jan CLI support - **added** — Add agent mode toggle and sandbox config isolation for OpenClaw - **added** — Add model provider options - **added** — Add dir=auto for bidirectional text support - **changed** — Jan manages Context Allocation dynamically - **changed** — Update Jan Docs - **changed** — Polish the initialize flow - **changed** — Update starter screen and download management UI - **changed** — Simplify mlx-server backend with new ChatSession update - **changed** — Better setting organizations - **changed** — Recommend small model - **fixed** — Duplicate url manifest flatpak - **fixed** — Conditional show message toolbar - **fixed** — Adjustment dragable app window - **fixed** — Thinking render with long content - **fixed** — Sorting provider name - **fixed** — MLX import dialog uses folder picker instead of file picker - **fixed** — Replace AbortSignal.any with manual polyfill for WebKit compatibility - **fixed** — Navigation back from detail hub page ##### What's Changed * Sync Release/v0.7.7 to main by @louis-jan in https://github.com/janhq/jan/pull/7504 * docs: changelog for v0.7.7 by @Vanalite in https://github.com/janhq/jan/pull/7507 * chore: update flatpak manifest v0.7.7 by @Minh141120 in https://github.com/janhq/jan/pull/7510 * chore: update landing page app screenshots by @urmauur in https://github.com/janhq/jan/pull/7511 * fix: duplicate url manifest flatpak by @Minh141120 in https://github.com/janhq/jan/pull/7519 * Sync new change from Release/v0.7.7 to main by @louis-jan in https://github.com/janhq/jan/pull/7518 * ci: add condition for nightly external by @Minh141120 in https://github.com/janhq/jan/pull/7523 * fix: conditional show message toolbar by @urmauur in https://github.com/janhq/jan/pull/7524 * fix: adjustment dragable app window by @urmauur in https://github.com/janhq/jan/pull/7521 * fix: incorrect links in docs by @statxc in https://github.com/janhq/jan/pull/7517 * fix: thinking render with long content by @urmauur in https://github.com/janhq/jan/pull/7527 * fix: sorting provider name by @urmauur in https://github.com/janhq/jan/pull/7526 * fix: simplify mlx-server backend with new ChatSession update by @louis-jan in https://github.com/janhq/jan/pull/7538 * fix: gpt-oss disallows quantized cache types due to its use of attention sinks by @louis-jan in https://github.com/janhq/jan/pull/7541 * fix: MLX import dialog uses folder picker instead of file picker by @louis-jan in https://github.com/janhq/jan/pull/7554 * fix: long text error apikey by @urmauur in https://github.com/janhq/jan/pull/7557 * fix: replace AbortSignal.any with manual polyfill for WebKit compatibility by @since-2017-hub in https://github.com/janhq/jan/pull/7552 * fix: navigation back from detail hub page by @urmauur in https://github.com/janhq/jan/pull/7569 * fix: use libappindicator3-1 for deb package to fix Ubuntu 22.04 install conflict by @dev-miro26 in https://github.com/janhq/jan/pull/7570 * feat: add xAI provider by @louis-jan in https://github.com/janhq/jan/pull/7551 * fix: add dir=auto for bidirectional text support by @sesajad in https://github.com/janhq/jan/pull/7542 * fix: hide capability for mlx model by @urmauur in https://github.com/janhq/jan/pull/7579 * docs: add Jan-Code-4B, Jan-v3-4B, and Jan-v2-VL model pages by @louis-jan in https://github.com/janhq/jan/pull/7598 * fix: add back linux native top panel tauri by @urmauur in https://github.com/janhq/jan/pull/7580 * feat: openclaw integration by @Vanalite in https://github.com/janhq/jan/pull/7583 * fix: disconnect channel and add spinner by @Vanalite in https://github.com/janhq/jan/pull/7599 * feat: polish the initialize flow by @Vanalite in https://github.com/janhq/jan/pull/7602 * feat: Jan CLI support by @louis-jan in https://github.com/janhq/jan/pull/7576 * fix: update flash attention handling in ArgumentBuilder by @dev-miro26 in https://github.com/janhq/jan/pull/7565 * fix: overlap z-index model setting and dropdown model provider by @urmauur in https://github.com/janhq/jan/pull/7600 * enhancement: better setting organizations by @urmauur in https://github.com/janhq/jan/pull/7601 * feat: Jan manages Context Allocation dynamically by @louis-jan in https://github.com/janhq/jan/pull/7609 * fix: disable auto fit setting by default by @louis-jan in https://github.com/janhq/jan/pull/7610 * fix: scrollable dropdown assistant by @urmauur in https://github.com/janhq/jan/pull/7607 * enhancement: recommend small model by @urmauur in https://github.com/janhq/jan/pull/7606 * fix: correct model eval chart by @louis-jan in https://github.com/janhq/jan/pull/7614 * fix(tauri): handle non-UTF-8 paths in default_data_folder_path by @eren-karakus0 in https://github.com/janhq/jan/pull/7577 * enhancement: add model provider options by @urmauur in https://github.com/janhq/jan/pull/7611 * fix: allow different images with same filename by deduplicating on co… by @mykh-hailo in https://github.com/janhq/j _[Truncated at 4000 characters — full notes: https://github.com/janhq/jan/releases/tag/v0.7.8]_ ### v0.7.7 — 0.7.7 - Date: 2026-02-11 - Version: v0.7.7 - Original notes: https://github.com/janhq/jan/releases/tag/v0.7.7 - Permalink: https://whatsnew.fyi/product/jan/releases/v0.7.7 - Labels: Platforms: Desktop - **fixed** — Theme set loop on Windows - **fixed** — Left panel full height - **fixed** — Throttle renderer to reduce pressure to the thread - **added** — Prompt to download Jan VL model for vision capability - **fixed** — Inconsistent code block formatting in Markdown rendering - **fixed** — Accept x-api-key header in place of custom header - **fixed** — Display embedding models in settings with visual distinction - **fixed** — Update API docs - **changed** — Default instruction update - **added** — Streamable HTTP Exa default search MCP - **changed** — Revamp layout - **fixed** — Minor badge capabilities - **fixed** — Dynamically generate OpenAPI spec with actual server host and port - **fixed** — Prefill value when rename thread - **added** — Delete all thread inside projects - **fixed** — Inline math latex - **added** — New backend - MLX - **fixed** — Chat performance rerender issue - **fixed** — Prompt cache - cannot infer the shape of an empty array - **added** — Implement fit settings in llamacpp extension and overhaul argument builder tests ##### What's Changed * fix: theme set loop on Windows by @louis-jan in https://github.com/janhq/jan/pull/7407 * fix: left panel full height by @urmauur in https://github.com/janhq/jan/pull/7409 * fix: throttle renderer to reduce pressure to the thread by @louis-jan in https://github.com/janhq/jan/pull/7408 * feat: prompt to download Jan VL model for vision capability by @Vanalite in https://github.com/janhq/jan/pull/7410 * docs: changelog for v0.7.6 by @Vanalite in https://github.com/janhq/jan/pull/7418 * chore: update manifest for flatpak jan 0.7.6 by @Minh141120 in https://github.com/janhq/jan/pull/7419 * chore: onboard new user with q4 k xl by @louis-jan in https://github.com/janhq/jan/pull/7420 * Fix: inconsistent code block formatting in Markdown rendering by @bytrangle in https://github.com/janhq/jan/pull/7049 * fix: accept x-api-key header in place of custom header by @louis-jan in https://github.com/janhq/jan/pull/7421 * fix: Display embedding models in settings with visual distinction by @bittoby in https://github.com/janhq/jan/pull/7393 * chore: sync release branch for merging by @Vanalite in https://github.com/janhq/jan/pull/7425 * chore: merge release branch by @Vanalite in https://github.com/janhq/jan/pull/7424 * fix: update api docs by @louis-jan in https://github.com/janhq/jan/pull/7426 * fix: default instruction update by @louis-jan in https://github.com/janhq/jan/pull/7427 * feat: streamable http exa default search mcp by @louis-jan in https://github.com/janhq/jan/pull/7428 * chore: disable auto updater for flatpak build by @hiento09 in https://github.com/janhq/jan/pull/7432 * ci: move config yarn makefile to ci by @Minh141120 in https://github.com/janhq/jan/pull/7433 * ui: revamp layout by @urmauur in https://github.com/janhq/jan/pull/7431 * fix: minor badge capabilities by @urmauur in https://github.com/janhq/jan/pull/7435 * fix: dynamically generate OpenAPI spec with actual server host and port by @Eruis2579 in https://github.com/janhq/jan/pull/7434 * fix: take back refresh and add model remote provider by @urmauur in https://github.com/janhq/jan/pull/7436 * fix: prefill value when rename thread by @urmauur in https://github.com/janhq/jan/pull/7437 * chore: delete all thread by @urmauur in https://github.com/janhq/jan/pull/7447 * fix: edit message with files attachment by @urmauur in https://github.com/janhq/jan/pull/7446 * chore: update font from local instead cdn by @urmauur in https://github.com/janhq/jan/pull/7449 * chore: avoid warning dialog components by @urmauur in https://github.com/janhq/jan/pull/7450 * chore: assistant assign into projects by @urmauur in https://github.com/janhq/jan/pull/7441 * feat: add delete all thread inside projects by @urmauur in https://github.com/janhq/jan/pull/7453 * fix: inline math latex by @urmauur in https://github.com/janhq/jan/pull/7454 * fix: hover menu dots for delete all chat by @urmauur in https://github.com/janhq/jan/pull/7455 * fix: notarize mlx bin by @Minh141120 in https://github.com/janhq/jan/pull/7458 * feat: add new backend - MLX by @louis-jan in https://github.com/janhq/jan/pull/7459 * chore: adjustment drag area by @urmauur in https://github.com/janhq/jan/pull/7464 * fix: chat performance rerender issue by @urmauur in https://github.com/janhq/jan/pull/7465 * fix: issue descendant element by @urmauur in https://github.com/janhq/jan/pull/7466 * fix: hub screen scroll and loader by @urmauur in https://github.com/janhq/jan/pull/7467 * fix: prompt cache - cannot infer the shape of an empty array by @louis-jan in https://github.com/janhq/jan/pull/7470 * chore: refactor modelhub by @hiento09 in https://github.com/janhq/jan/pull/7468 * chore: base layout windows by @urmauur in https://github.com/janhq/jan/pull/7471 * feat: implement fit settings in llamacpp extension and overhaul argument builder tests by @qnixsynapse in https://github.com/janhq/jan/pull/7442 * chore: update copy for setup screen by @urmauur in https:/ _[Truncated at 4000 characters — full notes: https://github.com/janhq/jan/releases/tag/v0.7.7]_ ### v0.7.6 — 0.7.6 - Date: 2026-01-27 - Version: v0.7.6 - Original notes: https://github.com/janhq/jan/releases/tag/v0.7.6 - Permalink: https://whatsnew.fyi/product/jan/releases/v0.7.6 - Labels: Platforms: Desktop - **added** — Combine attachments icon - **added** — Flatpak disable auto updater - **added** — Polish metainfo translation - **added** — Portable Windows build - **added** — Chat enhancement - **added** — Persist token usage into message metadata - **added** — Search dialog - **added** — Edit message and delete functionality - **added** — Enable proxy to llama.cpp for Anthropic Messages API support - **changed** — Update Russian localization - **changed** — Complete Czech translations - **changed** — Move llama.cpp config handling to Rust - **changed** — Improve sidebar UI with distinct sections and better empty states - **changed** — Migrate llamacpp backend logic to rust plugin - **changed** — Improve toast messages and errors - **fixed** — Cstdint patch flatpak - **fixed** — Flatpak appstream metadata validation errors - **fixed** — Improve Hub download button UX for installed models - **fixed** — Publisher name on Windows - **fixed** — Prevent CUDA backend install failure by removing invalid fs.mkdir args ##### What's Changed * docs: update homepage SEO meta title and description by @eckartal in https://github.com/janhq/jan/pull/7024 * chore: update flatpak manifest by @Minh141120 in https://github.com/janhq/jan/pull/7025 * blog: optimize "run AI locally" blog post for SEO by @eckartal in https://github.com/janhq/jan/pull/7028 * feat: combine attachments icon by @urmauur in https://github.com/janhq/jan/pull/6998 * fix: cstdint patch flatpak by @Minh141120 in https://github.com/janhq/jan/pull/7032 * Fix/appstream caption by @Minh141120 in https://github.com/janhq/jan/pull/7033 * feat: Update Russian localization by @Kuzmich55 in https://github.com/janhq/jan/pull/7039 * SEO: Improve offline-chatgpt-alternative content by @eckartal in https://github.com/janhq/jan/pull/7040 * feat: flatpak disable auto updater by @Minh141120 in https://github.com/janhq/jan/pull/7054 * Update flatpak metainfo by @eckartal in https://github.com/janhq/jan/pull/7064 * fix: flatpak appstream metadata validation errors by @Minh141120 in https://github.com/janhq/jan/pull/7067 * docs: add Microsoft Store and Flathub badges by @qnixsynapse in https://github.com/janhq/jan/pull/7068 * feat: add Polish metainfo translation by @0rzech in https://github.com/janhq/jan/pull/7050 * fix: improve Hub download button UX for installed models by @Lokimorty in https://github.com/janhq/jan/pull/7060 * Resolve Clippy Warnings and Apply Consistent Code Formatting by @github-roushan in https://github.com/janhq/jan/pull/7010 * feat: Update Russian localization by @Kuzmich55 in https://github.com/janhq/jan/pull/7082 * feat: complete Czech translations by @MichalZem in https://github.com/janhq/jan/pull/7081 * Update metadata for ai.jan.Jan by @razzeee in https://github.com/janhq/jan/pull/7045 * chore: update flatpak release 0.7.4 by @Minh141120 in https://github.com/janhq/jan/pull/7117 * chore: add change log doc for v0.7.4 by @Vanalite in https://github.com/janhq/jan/pull/7110 * refactor: clean up JoyRide by @louis-jan in https://github.com/janhq/jan/pull/7073 * chore: update flatpak manifest for v0.7.5 by @Minh141120 in https://github.com/janhq/jan/pull/7133 * chore: add change log doc for v0.7.5 by @Vanalite in https://github.com/janhq/jan/pull/7132 * chore: resolve conflicts between dev and 0.7.5 release branch by @Vanalite in https://github.com/janhq/jan/pull/7135 * chore: merge release 0.7.5 into dev branch by @Vanalite in https://github.com/janhq/jan/pull/7134 * refactor: move llama.cpp config handling to Rust by @qnixsynapse in https://github.com/janhq/jan/pull/7047 * fix: publisher name on windows by @Minh141120 in https://github.com/janhq/jan/pull/7144 * feat: Update Russian localization by @Kuzmich55 in https://github.com/janhq/jan/pull/7149 * feat: Improve toast messages and errors by @github-roushan in https://github.com/janhq/jan/pull/6892 * feat: improve sidebar UI with distinct sections and better empty states by @github-roushan in https://github.com/janhq/jan/pull/6745 * fix: prevent CUDA backend install failure by removing invalid fs.mkdir args by @qnixsynapse in https://github.com/janhq/jan/pull/7162 * fix: responsive button action hero landing page by @urmauur in https://github.com/janhq/jan/pull/7274 * refactor: jan web ci to use main branch and deprecate old workflows by @Minh141120 in https://github.com/janhq/jan/pull/7332 * refactor: migrate llamacpp backend logic to rust plugin by @qnixsynapse in https://github.com/janhq/jan/pull/7171 * feat: add portable windows by @Minh141120 in https://github.com/janhq/jan/pull/7350 * feat: chat enhancement by @louis-jan in https://github.com/janhq/jan/pull/7361 * feat: persist token usage into message metadata by @louis-jan in https://github.com/janhq/jan/pull/7366 * ci: update matrix test on windows by @Minh141120 in https://github.com/janhq/jan/pull/7367 * fix: tools availability and model settings by @louis-jan in https://github.com/janhq/jan/pull/7368 * fix: file upload fu _[Truncated at 4000 characters — full notes: https://github.com/janhq/jan/releases/tag/v0.7.6]_ ### v0.7.5 — 0.7.5 - Date: 2025-12-08 - Version: v0.7.5 - Original notes: https://github.com/janhq/jan/releases/tag/v0.7.5 - Permalink: https://whatsnew.fyi/product/jan/releases/v0.7.5 - Labels: Platforms: Desktop - **fixed** — Model import issue on Windows - **fixed** — Users couldn't import the mmproj vision model if the base name didn't match - **fixed** — CUDA backend installation and migration issues - **fixed** — Kill mcp server on app termination ##### What's Changed * fix: model import issue on Windows by @louis-jan in https://github.com/janhq/jan/pull/7124 * fix: users couldn’t import the mmproj vision model if the base name didn’t match. by @louis-jan in https://github.com/janhq/jan/pull/7127 * fix: correct CUDA backend installation and migration issues by @qnixsynapse in https://github.com/janhq/jan/pull/7130 * fix: kill mcp server on app termination by @Vanalite in https://github.com/janhq/jan/pull/7131 **Full Changelog**: https://github.com/janhq/jan/compare/v0.7.4...v0.7.5