SageChat

This update improves SageChat Pro subscriptions and polish. Claude Fable 5 has returned. Claude Sonnet 5 is now available as an option. SUBSCRIPTIONS • Fixed an issue where monthly and annual Pro plans could fail to load in TestFlight and the App Store build • Paywall now shows monthly and annual Pro plans with clear, direct pricing • Removed outdated free-trial messaging from the upgrade screen • Weekly plan is no longer offered for new subscriptions IMPROVEMENTS • More reliable Pro upgrade and restore flows • Clearer subscription pricing and plan selection in the paywall Thank you for using SageChat. If Pro plans still do not appear after updating, try again on a fresh install or contact support at sagekeylabs.ai.

This update improves SageChat Pro AI transcription and summary reliability. • Fixed hosted audio transcription retry and rate-limit handling. • Fixed AI summary generation so SageChat Pro AI summaries route through the hosted provider correctly. • Added clearer messages when hosted chat or transcription reaches minute, daily, or monthly fair-use limits. • Improved background transcription queue behavior and reduced repeated retry loops. • Reduced startup audio-session and app-group preference warnings for a smoother launch experience.

SageChat 2.2.7 makes the app easier to start and clearer to upgrade. - New onboarding starts free users with local/on-device AI instead of requiring provider API keys. - SageChat Pro now clearly unlocks hosted SageChat Pro AI access, voice workflows, web search, full history, and advanced controls. - Provider API keys are now optional advanced setup for users who want direct provider control. - Improved subscription messaging, privacy copy, and feature gating so Free vs Pro access is easier to understand. - Fixed first-launch onboarding behavior when reinstalling or testing with restored subscription entitlements.

Claude Fable 5 — Temporarily Unavailable Fable 5 has been blocked by the government and is temporarily unavailable. We've updated the app to reflect this: • Fable 5 can't be selected until the block is lifted • Claude Opus 4.8, Opus 4.7, Sonnet 4.6, and Haiku 4.5 remain fully available • All other Claude models continue to work in single-model chat and Multi-Model compare mode • Bring your own Anthropic API key for direct provider billing Plus the usual round of stability improvements and refinements. Thanks for using SageChat! Questions or feedback? We'd love to hear from you.

Free AI for everyone, every day. • 5 free SageChat Pro AI messages daily — no account, no API keys, just open and chat • Smarter first launch: SageChat now starts on an AI that's ready to answer immediately • Free trial pricing shown right on the upgrade screen when available • Cleaner Settings — advanced controls (custom prompts, model tuning) now live in a tidy Advanced section • Conversation list polish: provider icons on every chat, delete confirmation, smoother scrolling • Optional reminder when your free messages refresh each day • Every AI provider now has its own distinct icon throughout the app

NEW: CLAUDE FABLE 5 Anthropic's most powerful model has arrived — a new tier above Opus. • Pick Claude Fable 5 from the model selector in any chat • Available alongside Claude Opus 4.8, Opus 4.7, Sonnet 4.6, and Haiku 4.5 • Works in single-model chat and Multi-Model compare mode • Use your own Anthropic API key for direct provider billing SAGECHAT PRO AI — INCLUDED WITH PRO • Start chatting with hosted AI the moment you subscribe — no API key setup required • Fair-use access included in every Pro plan ALSO IN THIS UPDATE • Gemini 3.5 Flash is now the default Google Gemini model • Smarter unified chat context for more coherent long conversations • Reliability and performance improvements throughout As always, free on-device AI (Apple Intelligence, Qwen, and Gemma) keeps private chats on your iPhone — no account, no cloud, no API key.

SageChat 2.2.2 makes Pro easier to start using. - SageChat Pro now includes hosted AI for subscribers, with no provider API key setup required. - Added hosted transcription support for Pro voice workflows. - Improved onboarding and provider selection for SageChat Pro AI, on-device AI, and optional Bring Your Own Key providers. - Removed the weekly subscription option from the in-app purchase flow. - Updated privacy disclosures and review readiness for hosted AI features.

Claude Opus 4.8 is now available in SageChat. • Added Claude Opus 4.8 as the latest Anthropic flagship model • Made Opus 4.8 the default Claude option for new chats • Updated Claude model selection, onboarding, and in-app labels • Improved test coverage for Claude model availability and web search support Bring your own Anthropic API key and start using Claude Opus 4.8 directly in SageChat.

Version 2.1.2 - Added Gemini 3.5 Flash as the new default Gemini model. - Added GPT-5.5 Instant for fast OpenAI chats with your own API key. - Improved Share Sheet launch behavior so shared text and links open more reliably in SageChat. - Refined model selection and provider defaults. - Included reliability fixes and small polish throughout the app.

- Added GPT-5.5 Instant as a new OpenAI model option. - Improved trial setup reminders so they dismiss properly after snoozing. - Fixed the privacy consent flow so your first prompt is preserved and sent after you agree. - Updated model labels, setup copy, and related stability checks.

New: Share Sheet Import - Send links, text, images, PDFs, and files to SageChat from the iOS share sheet. - Shared content opens as a draft so you can review it before sending. - Fixed the handoff from the share sheet into SageChat.

What’s New • Added support for Grok 4.3 as a new model option • Fixed chat view spacing so messages and follow-up suggestions stay within the screen margins • Fixed a few smaller bugs to improve stability and polish • Made general performance and reliability improvements

Better voice recording, especially when your signal is weak. • Fixed a bug in the audio recording feature. • Tap Stop & Send and your recording shows up in your chat right away as "Transcribing…" — no more wondering if it saved. • Transcripts fill in as soon as they're ready, and patiently wait for a better connection if needed. • Fixed the false "Too Quiet" message on quick recordings. • Clearer messages when something does go wrong.

We fixed some bugs, cleaned up a few rough edges, and made SageChat a little more reliable. Nothing flashy, just the good kind of boring.

- Added GPT-5.5 support as the new default OpenAI model. - Added DeepSeek V4 support with DeepSeek V4 Flash and DeepSeek V4 Pro models. - GPT-5.5 and DeepSeek are now available across chat, model selection, onboarding, provider settings, and multi-model comparison. - Updated provider setup and privacy information for DeepSeek API keys. - Refreshed app copy to reflect support for 9 AI providers and the latest OpenAI model. - Improved provider display names, icons, and model labels throughout the app.

SageChat now supports Moonshot Kimi. - Added Moonshot Kimi as a new AI provider - Added Kimi K2.6, Moonshot’s latest model, with Kimi K2.5 also available - Kimi supports streaming chat, image input, and PDF text fallback - Added Kimi setup to onboarding, Settings, model selection, and privacy details - Updated multi-model chat, follow-up questions, and provider access logic for Kimi - Refined provider labels and subscription copy across the app - Improved provider tests and reliability checks

Updated screenshot graphics for the App Store.

A calmer, more confident SageChat. Rebuilt from the ground up around a single warm accent and cleaner typography, so your conversations — not the chrome — take center stage. What's new • Every AI, in one place. Switch between Claude, GPT, Gemini, and Grok right from the composer. Each reply shows which model answered. • Reimagined voice recording. Record inline without leaving your chat, or expand to Focus mode when an idea deserves the whole screen. A new review step lets you fix typos before you send. • Web search for Claude. Opus 4.7, Sonnet 4.6, and Haiku 4.5 can now pull live information from the web. • Smarter speech detection. Silent recordings no longer produce phantom transcripts. Plus: a rewritten Settings screen, faster starts, and reliability fixes for first-launch recordings.

• Claude Opus 4.7 — Anthropic's newest flagship model is now available alongside Opus 4.6, Sonnet 4.6, and Haiku 4.5; works in single-model chat and Multi-Model compare mode • Describe any photo — attach an image and ask GPT-5.4, Claude, Gemini, or Grok to identify objects, read signs, summarize screenshots, or explain what's happening • Read full documents — attach a PDF and get native understanding of tables, layout, and images on GPT-5.4, GPT-4o, Claude, and Gemini • Voice recordings anywhere in chat — record and insert voice messages at any point in a conversation, not just at the start; transcriptions happen automatically • Full-screen image viewer — tap any attached photo to view it at full resolution with zoom and share • Grok multi-agent mode fixed — xAI's collaborative multi-agent model now works reliably in both single chat and Multi-Model compare mode • Gemma Local — Google's open-weights Gemma model now runs entirely on-device, joining Qwen and Apple Intelligence as offline options • Smarter attachment routing — if your selected model can't read a file, SageChat tells you immediately and suggests one that can • Performance improvements for on-device inference and faster streaming across all cloud providers

NEW: Gemma 4 On-Device AI Run Google DeepMind's Gemma 4 models entirely on your device — no API key, no internet, complete privacy. • Gemma 4 E2B — lightweight and efficient (2B effective parameters) • Gemma 4 E4B — higher quality with 128K context window • Download once, use offline forever • Free for all users — no Pro subscription required Also improved: • Faster text rendering during on-device streaming • SageChat now supports 7 AI providers

One app. Every major AI — now with more ways to go Pro. - New weekly plan ($1.99/wk) — try Pro with less commitment - 14-day free trial on all Pro plans - Improved onboarding with quick-start tips for new users - Smarter review prompts — we'll ask for feedback at the right time - Updated pricing across all plans Already Pro? Thank you for your support. Enjoying SageChat? Leave a review — it helps more than you know.

Added support for GPT-5.4 Mini and GPT-5.4 Nano, giving you faster and lower-cost OpenAI options alongside GPT-5.4. We also improved model handling so saved selections stay consistent across updates, and added a more thoughtful in-app rating and feedback flow that appears only after meaningful use. Plus general polish and reliability improvements throughout SageChat.

Introducing SageChat Pro SageChat now offers Free and Pro tiers with a generous 7-day free trial on all Pro plans. Free tier includes: - Apple Intelligence & Qwen Local on-device AI — always free, no API key needed - Basic single-model chat - Last 10 conversations - Copy text export Pro tier unlocks everything: - All 6 AI providers: OpenAI (GPT-5.4), Anthropic Claude (Opus 4.6), Google Gemini (3.1 Pro), xAI Grok (4.20 Multi-Agent), plus Apple Intelligence & Qwen Local - Multi-model chat — compare responses from multiple AI providers side by side - Unlimited conversation history - Voice recording & transcription - Custom system prompts and personas - Full export (PDF, Markdown, share links) - Web search integration - iCloud sync across devices - Priority support & early access to new features Pro plans: $3.33/mo billed annually ($39.99/yr) or $5.99/mo. Both include a 7-day free trial. Qwen Local On-Device AI A second on-device AI option alongside Apple Intelligence. Qwen 3.5 models (0.8B, 2B, and 4B) run entirely on your device using Apple's MLX framework. Download once, then use offline with complete privacy. No API key or internet connection required. Available on both Free and Pro tiers. Smarter Feature Gating Pro-only features are now clearly marked throughout the app with lock indicators. Free users can browse all settings and see what's available before upgrading. Tapping any locked feature presents the upgrade screen with pricing and trial details. Improvements - Updated onboarding flow with Pro pricing, free trial info, and Qwen Local - Settings now display your subscription status with renewal/trial dates - Provider selection shows lock icons for cloud providers on the free tier - Subscription status syncs automatically to the Apple Watch companion app - Streamlined model selector gates cloud providers behind Pro subscription Your API keys are always stored locally on your device and never sent to SageChat servers. You pay AI providers directly with no markup — SageChat Pro simply unlocks access to use them.

Grok 4.20 Models - Grok 4.20 Multi Agent — coordinate multiple agents for complex tasks - Grok 4.20 Reasoning — advanced step-by-step reasoning - Grok 4.20 Non-Reasoning — fast, direct responses Qwen 3.5 On-Device - Run Qwen 3.5 locally on your device with no internet required - Available in 0.8B, 2B, and 4B model sizes (4-bit quantized) - No API key needed — fully private, on-device inference Smarter Follow-Up Questions - Follow-up questions are now generated using OpenAI (gpt-4o-mini) when available for faster, more reliable suggestions - When no OpenAI key is configured, follow-ups are generated by your currently selected provider - Added Grok support for follow-up question generation - Added Qwen Local on-device support for follow-up question generation — works fully offline Bug Fixes & Improvements - Improved provider fallback logic when generating follow-up suggestions - Apple Intelligence now serves as a last-resort fallback for follow-ups on supported devices

- GPT-5.4 — Access OpenAI's latest and most capable model, now set as the default - GPT-5.3 Instant — A faster, lighter GPT-5.3 variant for quick responses - Redesigned Chat Experience — Completely refactored chat interface inspired by modern AI apps: cleaner message layout, streamlined input bar, and improved conversation flow - Real-Time Streaming — Live token-by-token responses from all cloud providers (OpenAI, Claude, Gemini, Grok) for a more natural, responsive feel - Bug fixes and performance improvements