Practical overview
When it fits
Use Box when you want to explore local AI models or workloads on an Android device, especially when offline inference is a priority and the device meets the project’s listed requirements.
Who it is for
- Android users who want to run language models locally rather than use an internet-connected inference service.
- People testing .litertlm or GGUF models on a phone.
- Users interested in other documented on-device workloads, including image generation and speech recognition.
Documented uses
- Import a local .litertlm or GGUF model and use it in AI Chat.
- Choose CPU, GPU or NPU as the inference option during model import.
- Explore documented local AI tasks beyond chat, including image generation, speech recognition and document analysis.
Requirements & limits
The project README provides Android APK downloads through GitHub Releases and installation through Obtainium. It lists Android 14+ as a requirement.
- The app requires Android 14+ according to the project README.
- A typical quantised language model needs about 4 GB of free storage; the listed RAM requirement is 6 GB, or 8 GB for Bonsai image generation.
- LiteRT supports .litertlm and .task models; llama.cpp supports GGUF models.
- The import flow offers CPU, GPU or NPU choices, but device support depends on the hardware.
Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.
Import a local model into Box
Load a .litertlm or GGUF model file into Box and make it available in AI Chat.
Before you start
- An Android 14+ device.
- About 4 GB of free storage for a typical quantised language model.
- A .litertlm or GGUF model file on the device, or copied there from another location.
Steps
Copy the model file to the device, such as in Downloads or over USB.
Open Box and select Model Manager from the drawer.
Tap Import and choose the model file.
Set a display name and select CPU, GPU or NPU.
Open AI Chat; the imported model should appear there.
Expected outcome
The imported model is listed in AI Chat for local use.
Guide limitations
- The README lists 6 GB of RAM as a requirement; 8 GB is listed for Bonsai image generation.
- The project’s requirements and steps do not guarantee that every model or compute option will work on every device.
Compare options for Model inference and serving
These projects document a common task and share a project role. This is a source-based comparison, not an execution benchmark or a guarantee of interchangeability.
| Project | Documented purpose | Platform | Access and requirements |
|---|---|---|---|
| Box | Import a local .litertlm or GGUF model and use it in AI Chat. | Android | The project README provides Android APK downloads through GitHub Releases and installation through Obtainium. It lists Android 14+ as a requirement. |
| Bodhi App | Discover and download GGUF models from Hugging Face that suit local hardware. | Linux, macOS, Self-hosted, Windows | The README documents a Mac desktop release and Docker images for CPU, CUDA, ROCm and Vulkan configurations. The project identifies itself as open source. |
| Offlinellm | Import a GGUF model and use it for local multi-turn chat. | Android | The documented release is an APK available from GitHub Releases. It requires Android 13 or later and arm64-v8a; the Vanilla APK requires users to supply and import their own GGUF model. |
Sources & review
Source record details
- Source availability
- Active
- Sources checked
- 16 October 2026
- Original website title
- jegly/Box · Official homepage title · retained in its source language