On-Device models are here. Run AI models directly on your iPhone and iPad. Share them across your local network if you want to. • On-Device models: Powered by MLX for the best performance on Apple Silicon, with support for all of Reins' features. • Local network sharing: Expose models on your iPhone or iPad over an Ollama- or OpenAI-compatible endpoint for other devices to use. • Download any MLX model: Search for a specific model and download it even if it's not listed in the library. • Tool calling is now available for MiniCPM5 on-device models. • The LFM 2.5 2.6B on-device model now works. It previously failed to run. • Bug fixes and improvements.

On-Device models are here. Run AI models directly on your iPhone and iPad. Share them across your local network if you want to. • On-Device models: Powered by MLX for the best performance on Apple Silicon, with support for all of Reins' features. • Local network sharing: Expose models on your iPhone or iPad over an Ollama- or OpenAI-compatible endpoint for other devices to use. • Download any MLX model: Search for a specific model and download it even if it's not listed in the library. • Fixed an issue where some on-device models' thinking process could appear in the message content instead of the thinking block. • Bug fixes and improvements.

On-Device models are here. Run AI models directly on your iPhone and iPad. Share them across your local network if you want to. • On-Device models: Powered by MLX for the best performance on Apple Silicon, with support for all of Reins' features. • Local network sharing: Expose models on your iPhone or iPad over an Ollama- or OpenAI-compatible endpoint for other devices to use. • Download any MLX model: Search for a specific model and download it even if it's not listed in the library. • Bug fixes and improvements.

LM Studio and OpenAI-compatible server support are here, along with a much better web tool and a lot of improvements. • LM Studio support: Connect the same way you connect to Ollama. Model capabilities, loaded instances, downloads with live progress. • OpenAI-compatible backend: Connect to any server that speaks the OpenAI API, like oMLX, llama.cpp, vLLM and more. • Smarter web tool: Targets what you actually asked for. More focused answers, up to 90% less context used. Now works on sites that used to fail. • Default chat configuration: Set a default system prompt and generation options once. Every new chat starts there. • Background downloads: Model pulls keep running when you leave the app with live progress. (Requires iOS/iPadOS 26+) • Load models: Load an installed model right from the Manage Models page with custom context size support for LM Studio. • Select multiple files and images at once, plus a fix for files incorrectly shown as unsupported. • Compose Mail tool on macOS: No longer iOS-only. • Send feedback from the app: New Support section in Settings. • Bug fixes and improvements.

LM Studio and OpenAI-compatible server support are here, along with a much better web tool and a lot of improvements. • LM Studio support: Connect the same way you connect to Ollama. Model capabilities, loaded instances, downloads with live progress. • OpenAI-compatible backend: Connect to any server that speaks the OpenAI API, like oMLX, llama.cpp, vLLM and more. • Smarter web tool: Targets what you actually asked for. More focused answers, up to 90% less context used. Now works on sites that used to fail. • Default chat configuration: Set a default system prompt and generation options once. Every new chat starts there. • Background downloads: Model pulls keep running when you leave the app with live progress. (Requires iOS/iPadOS 26+) • Load models: Load an installed model right from the Manage Models page with custom context size support for LM Studio. • Compose Mail tool on macOS: No longer iOS-only. • Send feedback from the app: New Support section in Settings. • Bug fixes and improvements.

LM Studio support is here, along with a much better web tool and a lot of improvements. • LM Studio support: Connect the same way you connect to Ollama. Model capabilities, loaded instances, downloads with live progress. • Smarter web tool: Targets what you actually asked for. More focused answers, up to 90% less context used. Also fixed on sites that previously failed. • Default chat configuration: Set a default system prompt and generation options once. Every new chat starts there. • Background downloads: Model pulls keep running when you leave the app with live progress. (Requires iOS/iPadOS 26+) • Compose Mail tool on macOS: No longer iOS-only. • Send feedback from the app: New Support section in Settings. • Bug fixes and improvements.

This update contains great improvements and minor bug fixes: - Background Generation: Reins now generates responses in the background on iOS/iPadOS 26.0+, so you don't have to keep the app open while waiting. - Improved System Prompt Context: Add current time or force a web_fetch after web_search (under System Prompt Context in Settings, disabled by default). - Better Code Blocks: Syntax highlighting and a floating copy button for code. - Docker Model Runner Support: Use DMR models via the Ollama-compatible endpoint. - Fixed an issue where the Download Models section wouldn't show any models. - General improvements and bug fixes.

This update contains great improvements and minor bug fixes: - Background Processing: Reins now generates responses in the background on iOS/iPadOS 26.0+, so you don't have to keep the app open while waiting. - Better Code Blocks: Syntax highlighting and a floating copy button for code. - Docker Model Runner Support: Use DMR models via the Ollama-compatible endpoint. - Improved Web Search: New toggle to force full page fetches (under System Prompt Context in Settings, disabled by default). Search results expanded to 5. - General improvements and bug fixes.

This update contains great improvements and minor bug fixes: - Background Processing: Reins now generates responses in the background, so you don't have to wait with the app open (Requires iOS or iPadOS 26.0+, enabled by default). - Improved Web Search: New toggle to force full page fetches, found under System Prompt Context in Settings (disabled by default). Search results expanded to 5. - New Chat Stats: See total duration, token count, and tokens/s. - Audio Model Support: See which models support audio. - General improvements and bug fixes

Reins V2 is here. Tools, web search, file attachments, and thinking mode. The full power of your models, finally in your pocket. What's new: - Models can now call tools and reason across multiple steps - Built-in web search — no API key needed - Attach PDFs, CSVs, and text files to any conversation - Browse, download, and manage models from the app - Manage multiple servers and switch between them instantly - Access Ollama Cloud or your server from anywhere - Thinking mode for models that reason before they answer - Branch any chat to explore alternative paths - Chat without leaving a trace with Temporary Chats - Continue cut-off responses right where they stopped - Export and import your conversations Plus a redesigned configure sheet, a wave of bug fixes and improvements under the hood.

This version brings bug fixes and improvements. Changes: - Fixed save as a new model functionality - Error messages are now clearer and more helpful - Various minor improvements

- Improved accessibility: all UI now respects system font size settings. - Fixed issue where thinking indicator would disappear unexpectedly. - Fixed issue where chat title appeared empty while generating.

Fixed an issue where the chat could scroll up automatically while generating a response. Everything should now stay steady while you chat!

Bug fixes

Some bug fixes.

Now, reasoning messages from compatible models are shown separately for better readability.

Improved DeepSeek-R1 support! Thoughts are now separated with a collapsible section for better readability.

Bug fixes and UI improvements.