Atomic Chat v1.1.154

🚀 New Features

• Speculative decoding stats in the UI: token speed footer now shows tok/sec plus draft token acceptance %

🔧 Improvements & Fixes

• Windows DLL bundling: all upstream DLLs now ship together, backend prefs for TurboQuant vs upstream no longer overwrite each other (#175)
• Resumable downloads: model downloads now survive interruptions, resuming from the last saved byte via HTTP Range requests with exponential backoff (1s to 32s) instead of starting over
• Faster, safer MLX loads: text only checkpoints route through proper model classification, with a dedicated lock so a slow or failed load no longer blocks other models
• Backend version checks on Windows no longer pop a console window
• 5+ stability/UI fixes across model loading, downloads, and token timing

🙏 Contributors

Thanks to @Vect0rM, @AlexFromAtomic, @TheLeaderDev, @dtorey-d for their contributions to this release!

Atomic Chat v1.1.151

🚀 New Features

• DFlash support for llama.cpp — drafts whole blocks of tokens in one pass instead of token-by-token, giving up to ~2-3.5x faster generation on supported Qwen models with zero quality loss
• OpenClaude integration — added to the Launch page, launches an open coding agent against your local API server in one click
• New startup toggles — "Preload last used model on startup" (General) and "Auto-start Local API Server" (Server Configuration), so you can tune startup speed and behavior to your workflow

🔧 Improvements & Fixes

• More reliable llama-server startup detection across versions (#166)
• Fewer backend/model hangs and crashes overall
• Fixed a Linux crash where desktop notifications could abort the app or silently fail to deliver
• Tray status sync now works on Windows, matching macOS (#167)
• Fixed NVIDIA NIM cloud provider — models generate responses correctly again (#165)
• Integrations nav item now animates on hover

🙏 Contributors

Thanks to @worthant, @yanalialiuk, @danyurkin, @Vect0rM, @dtorey-d, @Alias4D, @jadams777 and @StephenLReed for their contributions to this release!

Atomic Chat v1.1.140

🚀 New features

• Added support for the Laguna XS 2.1 33B model by Poolside via TurboQuant llama.cpp — stronger results on multilingual coding and terminal-style tasks.
• TurboQuant support for llama.cpp on Windows and Linux gives you a bigger context window with lower memory usage, great for running agents on local models.
• VS Code, JetBrains and Xcode can now connect straight to your local server, with Base URL, key and model copied over automatically.
• We added our own quantized GGUF and MLX versions of Atomic Chat models to the catalog.
• New integration lets you detect, configure, and launch Zed (the fast Rust-based editor with a native AI agent panel) and Poolside CLI (a terminal agent built on the open ACP protocol), both pointed at your local server.
• Atomic Chat now auto-detects models you already have downloaded from Ollama, LM Studio, HF cache and Unsloth, so you can launch them instead of downloading them again.
• Ollama is now a built-in provider, so cloud and local Ollama models both show up and launch right away.

🔧 Improvements & fixes

• Pin the upstream llama.cpp backend to b9702 and update related assets for improved backend compatibility.
• The inference server port is now configurable instead of hardcoded to 1337, and Atomic Chat falls back automatically if that port is already taken by another app (#115).
• Custom provider model listing is fixed — models from vLLM or a remote llama.cpp server were not showing up (#90).
• You can now drag the window around on macOS again (#98).
• Model load errors are clearer now, including specific messages when an architecture isn't supported.
• Sending a message to a failed model is blocked, with a red status shown instead.
• RAG no longer leaks project files into other chats and system prompts stay intact when tools are active.
• Manifest fetch and backend downloads no longer get stuck on Linux.
• The Windows agent installer is more reliable, with proxy support, a PATH refresh after install, GPU checks and clearer errors.
• The model hub got some polish: cleaner README rendering and your search and scroll position is kept when you go back.
• • Recognize -qat and -unquantized suffixes in normalizeBaseId so QAT-quantized MLX models (e.g., gemma-4-*-qat-4bit) resolve correctly and speculative-drafter selection works; includes new unit tests.
• Plus 40+ other stability and UI fixes.

🙏 Contributors

Thanks to @yanalialiuk, @Vect0rM, @cashcon57, @danyurkin, @Albert-Atomic, @dtorey-d, @nuroctane, @caretak3r and @shrisha108 for their contributions to this release!

Atomic Chat v1.1.137

🚀 New features

• Added support for the Laguna XS 2.1 33B model by Poolside via TurboQuant llama.cpp — stronger results on multilingual coding and terminal-style tasks.
• TurboQuant support for llama.cpp on Windows and Linux gives you a bigger context window with lower memory usage, great for running agents on local models.
• VS Code, JetBrains and Xcode can now connect straight to your local server, with Base URL, key and model copied over automatically.
• We added our own quantized GGUF and MLX versions of Atomic Chat models to the catalog.
• New integration lets you detect, configure, and launch Zed (the fast Rust-based editor with a native AI agent panel) and Poolside CLI (a terminal agent built on the open ACP protocol), both pointed at your local server.
• Atomic Chat now auto-detects models you already have downloaded from Ollama, LM Studio, HF cache and Unsloth, so you can launch them instead of downloading them again.
• Ollama is now a built-in provider, so cloud and local Ollama models both show up and launch right away.

🔧 Improvements & fixes

• The inference server port is now configurable instead of hardcoded to 1337, and Atomic Chat falls back automatically if that port is already taken by another app (#115).
• Custom provider model listing is fixed — models from vLLM or a remote llama.cpp server were not showing up (#90).
• You can now drag the window around on macOS again (#98).
• Model load errors are clearer now, including specific messages when an architecture isn't supported.
• Sending a message to a failed model is blocked, with a red status shown instead.
• RAG no longer leaks project files into other chats and system prompts stay intact when tools are active.
• Manifest fetch and backend downloads no longer get stuck on Linux.
• The Windows agent installer is more reliable, with proxy support, a PATH refresh after install, GPU checks and clearer errors.
• The model hub got some polish: cleaner README rendering and your search and scroll position is kept when you go back.
• Plus 40+ other stability and UI fixes.

🙏 Contributors

Thanks to @yanalialiuk, @Vect0rM, @cashcon57, @danyurkin, @Albert-Atomic, @dtorey-d, @nuroctane, @caretak3r and @shrisha108 for their contributions to this release!

Atomic Chat v1.1.134

🚀 New features

• TurboQuant support for llama.cpp on Windows and Linux gives you a bigger context window with lower memory usage, great for running agents on local models.
• VS Code, JetBrains and Xcode can now connect straight to your local server, with Base URL, key and model copied over automatically.
• We added our own quantized GGUF and MLX versions of Atomic Chat models to the catalog.
• New integration lets you detect, configure, and launch Zed (the fast Rust-based editor with a native AI agent panel) and Poolside CLI (a terminal agent built on the open ACP protocol), both pointed at your local server.
• Atomic Chat now auto-detects models you already have downloaded from Ollama, LM Studio, HF cache and Unsloth, so you can launch them instead of downloading them again.
• Ollama is now a built-in provider, so cloud and local Ollama models both show up and launch right away.

🔧 Improvements & fixes

• The inference server port is now configurable instead of hardcoded to 1337, and Atomic Chat falls back automatically if that port is already taken by another app (#115).
• Custom provider model listing is fixed — models from vLLM or a remote llama.cpp server were not showing up (#90).
• You can now drag the window around on macOS again (#98).
• Model load errors are clearer now, including specific messages when an architecture isn't supported.
• Sending a message to a failed model is blocked, with a red status shown instead.
• RAG no longer leaks project files into other chats and system prompts stay intact when tools are active.
• Manifest fetch and backend downloads no longer get stuck on Linux.
• The Windows agent installer is more reliable, with proxy support, a PATH refresh after install, GPU checks and clearer errors.
• The model hub got some polish: cleaner README rendering and your search and scroll position is kept when you go back.
• Plus 40+ other stability and UI fixes.

🙏 Contributors

Thanks to @yanalialiuk, @Vect0rM, @cashcon57, @danyurkin, @Albert-Atomic, @dtorey-d, @nuroctane, @caretak3r and @shrisha108 for their contributions to this release!

Atomic Chat v1.1.119

🚀 New features

• Audio input for omni and MLX models — record and play audio right in the chat
• The app now remembers window size and position across restarts
• Download a model during onboarding and you'll jump straight into chat, with progress still visible
• Smarter sampling: recommended defaults for Gemma 4 QAT models, while keeping any values you've customized
• MLX models now have a proper Download section on the Hub (name, MLX badge, file size, Download / New Chat)

🔧 Improvements & fixes

• Fixed CUDA and Vulkan auto-detection on Windows — the right GPU backend now gets picked and downloaded automatically
• More reliable backend downloads overall (shared / NAT / VPN networks, fewer GitHub rate-limit failures, offline fallback preserved)
• Better Metal GPU recovery: clearer out-of-memory messages and auto-reload after compute errors
• Fixed system prompt being dropped in regular local-model chats
• Clickable links in the model-policy error banner
• Quieter logs (no more misleading CORS warnings for non-browser clients)

🙏 Contributors

Thanks to @Vect0rM, @Albert-Atomic for their contributions to this release!

Atomic Chat v1.1.115

🚀 New Features

• Add Cline CLI and MiMo Code coding agents to the Integrations page for one-click running
• Support WebP image attachments
• Add drag-and-drop for projects in the sidebar
• Add a configurable maximum image size with automatic downscaling when images exceed the limit
• Add Korean (한국어) language support

🔧 Improvements & Fixes

• Migrate macOS autostart from LaunchAgent to AppleScript, with a one-time migration to preserve existing autostart preferences
• Add tiered backend fallback when a pinned backend is unavailable, returns a 404, or is unreachable: newest installed same-type backend → newest published family tag (download) → any installed backend as a last-resort CPU degrade, with corrected pins persisted for tiers 1–2; also clean up orphaned/incomplete backend installs automatically
• Resolve the latest published CUDA minor version dynamically in Windows build scripts and handle GitHub API rate limits gracefully, preventing 404s
• Map local model-server transport errors to 503 with a structured OpenAI-style error envelope (including Retry-After); wrap local-engine errors with typed codes (e.g. context_length_exceeded, insufficient_memory)
• Validate model/.mmproj presence and file size before loading to avoid confusing failures
• Raise the model-load readiness timeout to 30 minutes (1800s) so large models on slow storage aren't cut off prematurely
• Surface CPU_NO_AVX as an actionable load error and fail fast on x86 CPUs without AVX support
• Classify macOS Catalina Metal dyld symbol-not-found and MLX Metal OOM/unsupported-arch crashes as actionable unsupported-OS/unsupported-arch errors instead of generic crashes
• Resolve the embedding engine under the upstream name first (falling back to the legacy name) so attachment ingestion works with the upstream llama.cpp engine
• Sync the UI context window display with the server's actual llama.cpp context size
• Default exported filenames to the project name
• Honor the Windows llamacpp alias in the Hub's Downloaded filter so downloaded models show correctly
• Fall back to a free port automatically if port 1337 is already in use
• Refresh the Hub's Downloaded list after a model is deleted so the UI stays current
• Add 10+ smaller stability and UI fixes across Windows and macOS (hover/chevron clipping, per-row hover isolation, click/drag activation and related sidebar UX)

🙏 Contributors

Thanks to @Vect0rM and @Albert-Atomic for their contributions to this release!

Atomic Chat v1.1.114

🚀 New Features

• Added a "Download a model" shortcut to the model dropdown for quicker model downloads from the selector.
• On macOS, closing the window now sends the app to the tray instead of quitting so the app keeps running in the background.
• Model loading now shows clearer, backend-aware errors for unsupported or corrupt models so you get actionable diagnostics when a model fails to load.

🔧 Improvements & Fixes

• Search moved to the bottom of the sidebar navigation so the Search item lives at the end of the nav list.
• Repair invalid tool-call JSON emitted by coder models instead of crashing, preventing interruptions when models produce malformed tool-call output.
• Sidebar tree view polish: tightened the project chevron spacing and aligned per-conversation overflow ("...") buttons so tree items look consistent and overflow actions appear in the right place.
• Windows maximize fix: caption buttons now fill the very top edge when the window is maximized so close/minimize/maximize are easier to hit.
• macOS agent and backend fixes: prevented the upstream auto-upgrade from deleting unrelated turboquant backends, added recovery to restore missing turboquant backends from bundled resources, improved WSL CLI-agent detection, and added manual binary-path overrides for non-standard installs.
• Hub loads hardware info on startup so model fit badges reflect correct hardware compatibility immediately.
• macOS permission: added NSLocalNetworkUsageDescription so LAN model servers can be discovered and used on macOS.
• Prevented Windows caption controls from overlapping the HTML preview toolbar to avoid UI overlap issues.

🙏 Contributors

Thanks to @danyurkin, @Vect0rM, @Albert-Atomic for their contributions to this release!

Atomic Chat v1.1.112

🚀 New Features

• Global sampling settings — sampling parameters are now a single persisted global store (Zustand) instead of per-assistant, so sampling behaves consistently across the app and legacy on-disk data is preserved.
• Pause/resume model downloads — resumable GGUF downloads can be paused and resumed from the global Downloads popover with new UI controls and resume-state handling for better download control.

🔧 Improvements & Fixes

• Clearer model loading feedback — previously persistent "Failed to load" toasts are cleared on successful load and engine error codes are mapped to actionable titles/descriptions, with updated English and Russian locale messages.
• Windows GGUF validation fix — model path validation was enhanced to avoid issues with split GGUF files on Windows, improving model loading reliability.
• Download UI and robustness tweaks — download button keeps consistent height during download states and paused downloads no longer trigger error notifications or cleanup.

🙏 Contributors

Thanks to @Vect0rM for their contributions to this release!

Atomic Chat v1.1.108

🚀 New Features

• Add one-click Kilo, Pi, Goose and OpenHands on the integration page to quickly start coding agents
• Enable "Launch at startup" by default for all users
• Render inline image attachments directly inside chat bubbles so images appear in-thread

🔧 Improvements & Fixes

• Prevent non-MTP models from being bricked by gating the global llamacpp-upstream MTP flag on a per-model capability at load time.
• 5+ stability and UI fixes

🙏 Contributors

Thanks to @Vect0rM and @danyurkin for their contributions to this release!