Changelog

Every release, written down.

1.3.0

Latest release ·

1.3.0

Latest release
iOSAndroidWeb
  • Custom LLMs rebuilt: each model now uses its own protocol, detected automatically, with new setup, detail and Model Library screens for relays and local compute
  • Connect remote MCP servers and let models use their tools in chat, with permissions for each tool
  • Oriveo AI can now search the web and cite its sources, and you can choose its thinking effort on supported models
  • Sign in with Microsoft, on a redesigned sign-in screen
  • New Notifications page and an in-app feedback formiOS · Android
  • Redesigned Model Options and Advanced settings that show what each model and connection can do
  • PDFs and other files now reach OpenAI, Anthropic, Gemini and OpenRouter models as real files, so scanned pages and layouts stay readable
  • More natural, consistent wording in all 16 languages
  • Improved stability on iOS 18iOS
Read release notes

1.2.10

iOSAndroidWeb
  • Long sent messages now collapse into a preview; tap to expand
  • Faster Home, provider list and model picker
  • New split-screen layout on iPad and wide windows, with content centered and width-limitediOS
  • Improved reliability of chat history and conversation synciOS · Android
  • More buttons, switches and password fields support screen readers and show/hide
  • Fixed lingering English and awkward translations in several languages
Read release notes

1.2.9

iOSAndroidWeb
  • New app icon and logo
  • Redesigned Home with a cleaner input box, grouped chats and a new Notes cardiOS · Android
  • Time-of-day greetings that read naturally in every languageiOS · Android
  • Smoother typing; faster Home, search and model listsiOS · Android
  • Math in tables now displays correctlyiOS · Android
  • Provider region choices now save correctly
  • Replies and drafts are saved more reliablyiOS
  • Chat follows new replies and the keyboard again after you scrollAndroid
  • Sent text no longer reappears in the input box when typing with an input method (IME)Web
  • Collapsing and expanding the sidebar is now smoothWeb
  • Attached web pages (HTML) are now sent as clean textWeb
Read release notes

1.2.8

iOSAndroidWeb
  • Settings, connections and account changes now sync more reliably across devices
  • Subscription connections keep the correct sign-in method after cross-device restore
  • Interrupted backup replacement resumes safely without restoring deleted cloud data
  • Temporary free-model failures now show consistent, localized retry guidance
  • Large model catalogs remain available offline and handle big caches safelyAndroid · Web
  • Loading older messages better preserves your reading positioniOS
  • Fixed Android 14 startup, long messages, billing recovery and external linksAndroid
  • Multi-tab sign-out, sync recovery and older-browser chat are more reliableWeb
Read release notes

1.2.7

iOSAndroidWeb
  • NEW: Sign in with your Grok or ChatGPT subscription, no API key needed
  • NEW: Models can call tools, and every tool call is visible in the chat
  • Faster startup, scrolling, streaming answers and model search
  • Free models now prompt you to sign in instead of showing an empty quota
  • Attachments already sent are no longer re-uploaded every turn
  • More reliable backup restore and better overall stability
Read release notes

1.2.6

iOSAndroidWeb
  • Model controls now combine web search, reasoning and model behavior, adapting to each model
  • Connect Ollama, LM Studio, llama.cpp, vLLM or Open WebUI on your computer or local network
  • Custom LLM setup now auto-detects protocols and model catalogs
  • Select chat text to ask or quote it, with context saved after sending
  • See input, output and cache token details for each reply
Read release notes

1.2.5

iOSAndroidWeb
  • NEW: Library lets models find and cite details from the Notion pages you choose
  • Redesigned Library screens for sources, connections and monthly usage
  • NEW: Relay connects to AI services on your local network or private VPN
  • Streaming answers now settle smoothly instead of flashing at the endiOS · Android
  • Scrolling back through a long chat keeps your reading positioniOS
  • Oriveo AI and free model lists keep the order we publish
Read release notes

1.2.4

iOSAndroidWeb
  • NEW: A guided tour of Oriveo on first launchiOS · Android
  • NEW: Mistral joins as our 15th official provider
  • New installs start in dark mode, and your theme syncs across devices
  • Provider list shows balances and spending, including your Oriveo AI wallet
  • OpenRouter web sources now arrive together with the finished answeriOS · Web
  • Qwen image models are recognized and requested correctly
  • Softer ambient glow around your messagesiOS · Android
  • No more flicker when you send the first message in a new chatWeb
  • Visual polish and general stability improvements
Read release notes

1.2.3

iOSAndroidWeb
  • NEW: Set a display name for your account
  • AI thinking now renders with proper formatting
  • More reliable continue, stop, and regenerate with Oriveo AI
  • Smoother typing and long conversations, with less heatiOS · Android
  • OpenRouter web search keeps reasoning and sources intactAndroid · Web
  • Refreshed relay type picker with colorful cardsiOS · Android
  • Improved data reliability when switching accounts or signing out
  • General stability improvements
Read release notes

1.2.2

iOSAndroidWeb
  • Math formulas now stay intact when you copy selected chat text or save it to Notes
  • Simplified model lists by removing redundant default labels
  • Notes now render math formulasiOS
  • Restored Save to Notes in the text-selection menuAndroid
  • Oriveo AI messages now send without an extra confirmation dialogWeb
Read release notes

1.2.1

iOSAndroidWeb
  • Oriveo AI models now show their thinking as they work
  • Refreshed the app icon and provider logos
  • Deleting a provider now syncs reliably across your devices
  • The model you choose before signing in is now kept after you log in
  • Faster notes recall across your devicesiOS · Web
  • Smoother streaming - code blocks no longer jump and the device stays cooleriOS
  • Oriveo AI chats from your other devices now sync instead of being skippediOS
  • Fixed AI Balance top-ups occasionally being counted twiceAndroid
  • Fixed a crash when migrating older chat historyAndroid
Read release notes

1.2.0

iOSAndroidWeb
  • Meet Oriveo AI: chat with a curated model catalog without adding API keys
  • Added AI Balance top-ups, bonus credits, transaction history, and usage insights
  • Added weekly included usage, with automatic balance fallback when it runs out
  • Improved image generation across Grok, Qwen, MiniMax, Zhipu, and SiliconFlow
  • Fixed deleted conversations reappearing after cross-device sync
Read release notes

1.1.4

iOSAndroid
  • Fixed Gemini's Gemma models failing to reply when a thinking level was selected
  • Fixed conversations with large image/video/PDF/Office/file attachments becoming slow or laggyiOS
  • Fixed the message action toolbar occasionally appearing misaligned right after a reply finishediOS
  • Fixed Grok/Qwen/Zhipu/SiliconFlow image chats erroring out when a synced image hadn't downloaded locally yetiOS
  • Fixed OpenAI PDF/Office attachments failing to send in the same situationiOS
  • Fixed a crash opening chats with large image/video/PDF/Office/file attachments; existing oversized attachments are repaired automaticallyAndroid
  • Fixed the Grok logo being invisible in dark mode and the Together logo turning solid whiteAndroid
Read release notes

1.1.3

iOSAndroid
  • NEW: Conversation outlineIn long chats, tap the index along the right edge to jump straight to any of your questions - even across hundreds of messages. The dots move page by page and follow as you scroll.
  • Loading older messages now keeps your reading position rock-steadyiOS
  • Fixed rare crashes and scrolling glitches when navigating long chatsiOS
  • Fixed an occasional freeze on messages with special formattingiOS
  • Cleaned up the alerts shown when a quota runs out or a purchase is cancellediOS
  • Also fixed: the outline occasionally jumping to the wrong message.Android
Read release notes

1.1.2

iOSAndroid
  • Added server-driven model capability metadata for reasoning, web search, and image generation requests
  • Chat auto-recovers when a provider rejects an unsupported parameter, instead of failing
  • Fixed chat errors on newer Grok models
  • Fixed OpenAI reasoning summaries missing in Automatic mode
  • Fixed Zhipu web search never activating even when enablediOS
  • Fixed Qwen web search staying active after being turned offiOS
  • Image generation fixes for OpenRouter and GeminiiOS
  • Added Moonshot China (moonshot.cn) endpoint supportiOS
Read release notes

1.1.1

iOSAndroid
  • Fixed message-selection menu conflicts
  • Sync status: fixed a false "synced" indicator during degraded network periods, added auto-reconnect when stuck on "unavailable", and a clearer status icon
  • Fixed the message-completion action toolbar occasionally becoming misalignediOS
  • Fixed sending with the old model right after switching models mid-chatAndroid
  • Fixed abnormal message spacing after a failed sendAndroid
  • Fixed a layout glitch on the Providers/Settings tabsAndroid
Read release notes

Every published release is listed here. Older builds are still in the in-app changelog.

Follow development on GitHub