What problem it solves

Many AI applications send messages to remote servers, where they may be logged and analyzed outside the user's view. This project addresses that concern by running language models, conversations, voices, and tools on the device without requiring an account or internet connection.

Key features

• On-device language model chat — Run GGUF language models fully offline, including Gemma, Qwen, Phi, and Llama. Models can be downloaded from the app or added from Hugging Face.

  Source: README

• Voice input and speech — Dictate messages with on-device Whisper speech recognition and give assistants a voice through on-device neural voices or system voices.

  Source: README

• Personalized assistants — Create Assistant and Roleplay Pals with their own models, system prompts, personalities, and contextual parameters. Community Pals can also be discovered and installed through PalsHub.

  Source: README

• Tool-enabled conversations — Talents let capable Pals call built-in tools during a conversation. Supported tools include calculation, date and time, HTML rendering, and opt-in web search and page reading.

  Source: README

• Hardware-aware inference — The app supports CPU, GPU, and NPU inference paths with graceful fallback when a full backend is unavailable. It also provides device benchmarking for reviewing tokens per second and memory usage.

  Source: README

Quick start

Copied verbatim from the project's README.

git clone https://github.com/a-ghorbani/pocketpal-ai
cd pocketpal-ai

nvm use                       # match the pinned Node version
yarn install                  # install JS dependencies
(cd ios && pod install)       # iOS only

yarn start                    # Metro bundler
yarn ios                      # build + run on iOS simulator
yarn android                  # build + run on Android emulator

Who it may suit

This project may suit people who want private, offline AI conversations and voice features on phones or tablets without creating an account. It may also suit React Native developers interested in on-device inference, speech recognition, text-to-speech, and tool-enabled assistants.

Topics to discuss

• Which on-device models and quantizations work best across different phones without making specific performance claims?

• How should optional online features communicate their privacy implications and configuration requirements?

• What new Talents would make on-device assistants more useful without requiring an account or cloud service?

Written by ReadmeX AI from the project's README, source and public GitHub data; it may contain mistakes.

Read the original README https://github.com/a-ghorbani/pocketpal-ai#readme

ReadmeX opened this community for a trending project; it is not run by the project. Maintainers can claim it from the community page.