AppRecs review analysis
AppRecs rating 2.6. Trustworthiness 79 out of 100. Review manipulation risk 20 out of 100. Based on a review sample analyzed.
★★☆☆☆
2.6
AppRecs Rating
Ratings breakdown
5 star
20%
4 star
0%
3 star
40%
2 star
0%
1 star
40%
What to know
✓
Low review manipulation risk
20% review manipulation risk
✓
Credible reviews
79% trustworthiness score from analyzed reviews
⚠
Mixed user feedback
Average 2.6★ rating suggests room for improvement
About LLMConnect • AI Chat
Connect your own API keys for OpenAI, Anthropic, OpenRouter, Groq, Grok, Perplexity, DeepSeek, Gemini, Mistral, Replicate, or custom endpoints. Switch models on the fly.
Free plan available. Some features require a subscription. Provider usage fees may apply.
▸ All the models you need
Plug in keys for multiple providers and compare quality, speed, and cost in one place. Use Local Models on your device when offline.
▸ One unified chat
Folders, pins, powerful search, and swipe actions keep dozens of conversations organized. Create branch chats to explore ideas without losing the original thread. Lock the app, specific chats, or folders with Face ID or Touch ID. iCloud sync keeps chats and folders aligned across iPhone, iPad, and Mac. Export and import conversations in JSON.
▸ Write, translate, and summarize better
Draft professional emails, rewrite and proofread text, change tone, translate content, and create concise summaries of long text and PDFs.
▸ Built-in image generator
Create images and variations directly in the chat without switching apps.
▸ Prompt gallery + memory
Save reusable “/prompts” and give assistants ongoing context so they remember what matters in each thread.
▸ Privacy first and cost control
Your API keys are stored on device. Requests go directly from your device to the providers you select. You decide which models to use and how much to spend. Local Models run on device, even offline.
▸ Designed to move fast
New streaming and Markdown processors, Liquid Glass aesthetics, a new Black & White theme, redesigned message bubbles, a new typing indicator, UI refinements, and faster performance with improved stability.
▸ Background tasks
Close the app while a model is responding and get a notification when it’s done.
How it works:
1. Add your API keys or choose a Local Model.
2. Pick a provider and model, then start chatting instantly.
3. Organize conversations with folders or pins and reuse prompts to move faster.
4. Sync with other devices.
Perfect for
• Developers benchmarking multiple LLMs
• Content creators automating writing and design tasks
• Students and researchers who summarize and translate a lot of material
• Anyone who values cost transparency and data ownership
Subscription terms
Auto-renewable subscriptions. Payment is charged to your Apple ID at confirmation. The subscription renews unless you cancel at least 24 hours before the end of the current period. You can manage and cancel in Settings > Apple ID > Subscriptions. Any unused portion of a free trial is forfeited when you purchase a subscription.
Legal disclaimer
LLMConnect is an independent application and is not affiliated with, endorsed by, or sponsored by OpenAI, Anthropic, OpenRouter, Groq, Grok (xAI), Perplexity, DeepSeek, Google (Gemini), Mistral, Replicate, or any other third-party provider. All product names, logos, and brands are property of their respective owners. Use of these names is solely for describing compatibility.
Terms of Use: https://www.apple.com/legal/internet-services/itunes/dev/stdeula/
Tap to Rate:
Reviews for LLMConnect • AI Chat
forced tutorials are lame
Very buggy
Trying to use a custom provider. It connects and fetches models but always gives me Error: The operation couldn’t be completed. (CustomProvider error 2.) Lots of weirdness to the UI layout and navigation and gestures. This could be a solid four app if it worked right.
Toomanyapps!
Full featured and will be awesome someday!
Plagued with ridiculous bugs… Stuff that’s easy to fix. Like the menu button cannot be pressed. Yes, it can be but like it takes 10 presses. common Xcode problem. Also, RAG is not implemented correctly apparently. Would love to give them my money but too broken.