Anima - OnBrand Vibe Coding — Design-aware AI for modern product teams.
Design-aware AI for modern product teams.
Promoted
Maker
📌
🚀 Hey everyone! PocketPal got an update
PocketPal AI puts powerful language models right in your pocket — running entirely offline on your iPhone or iPad. No cloud, and your data stays private.
What PocketPal AI Offers: - On-device LLMs: Fully private, offline inference with models like Phi, Gemma, Qwen, and more. - Custom Pals: Build your own AI personas with unique personalities, prompts, and models. - Advanced controls: Adjust temperature, top-k, tokens/sec, and other parameters to fine-tune output. - Performance benchmarking: Compare SLMs and see how they run directly on your device. - Easy model management: Download, update, or swap models from Hugging Face or local files.
🔥 What's new in v1.10.15 (July 13, 2025): - New **"thinking control"**: For reasoning models like SmolLM3 easily toggle thinking on or off as needed.
Thanks for stopping by and we’re excited to hear your thoughts and grow with the community!
PocketPal lets you benchmark models and submit the results to this leaderboard. You can view a leaderboard of devices or explore how each model performs on specific phones in terms of tokens/sec, or do a 1-on-1 comparison between devices, if benchmark of similar models have been reported on them.
🚀 Hey everyone! PocketPal got an update
PocketPal AI puts powerful language models right in your pocket — running entirely offline on your iPhone or iPad. No cloud, and your data stays private.
What PocketPal AI Offers:
- On-device LLMs: Fully private, offline inference with models like Phi, Gemma, Qwen, and more.
- Custom Pals: Build your own AI personas with unique personalities, prompts, and models.
- Advanced controls: Adjust temperature, top-k, tokens/sec, and other parameters to fine-tune output.
- Performance benchmarking: Compare SLMs and see how they run directly on your device.
- Easy model management: Download, update, or swap models from Hugging Face or local files.
🔥 What's new in v1.10.15 (July 13, 2025):
- New **"thinking control"**: For reasoning models like SmolLM3 easily toggle thinking on or off as needed.
Thanks for stopping by and we’re excited to hear your thoughts and grow with the community!
Smoopit
Love the offline-first approach. How does performance compare between different model sizes on mobile hardware? @llm_ventures
@rachitmagon great question! I have something you may love: https://huggingface.co/spaces/a-ghorbani/ai-phone-leaderboard
PocketPal lets you benchmark models and submit the results to this leaderboard. You can view a leaderboard of devices or explore how each model performs on specific phones in terms of tokens/sec, or do a 1-on-1 comparison between devices, if benchmark of similar models have been reported on them.
Smoopit
@llm_ventures Thanks for sharing this.