Transparency & Attribution

Open Source Software & Credits

Tolliver is built on the work of researchers, developers, and open-source communities around the world. This page acknowledges the projects and tools that make Tolliver possible.

Tolliver currently uses several open-source technologies, AI models, and third-party services. Tolliver does not claim ownership of these tools — we are grateful to every researcher and developer who made them freely available.

AI Models & Offline Brain

Language & Intelligence

TinyLlama 1.1B Apache 2.0

Tolliver's offline AI model. Answers your questions when you're not connected to the internet. Copyright © Zhang Peiyuan et al., Singapore University of Technology and Design.

github.com/jzhang38/TinyLlama
llama.cpp MIT

Powers Tolliver's offline brain. Runs TinyLlama entirely on your phone with no internet connection. Copyright © Georgi Gerganov and contributors.

github.com/ggerganov/llama.cpp
Vision & Face Recognition

Seeing & Recognizing

MobileFaceNet MIT

The face recognition model Tolliver uses to identify family members on-device. Lightweight and built for mobile — it converts a face image into a compact mathematical signature without sending any data off your device. Copyright © 2019 syaringan357.

github.com/syaringan357/Android-MobileFaceNet​-MTCNN-FaceAntiSpoofing
ArcFace Verify License

The training method behind Tolliver’s face recognition. ArcFace loss trained MobileFaceNet to create more precise, better-separated face signatures — significantly improving real-world accuracy without changing the model size or speed. Based on “ArcFace: Additive Angular Margin Loss for Deep Face Recognition” (2019). Credit: Jiankang Deng, Jia Guo, Niannan Xue, Stefanos Zafeiriou — Imperial College London. License applies to the InsightFace codebase (github.com/deepinsight/insightface). Verify specific component licensing before release.

github.com/deepinsight/insightface
TensorFlow Lite Apache 2.0

Runs the face recognition model on-device. Keeps all face data on your phone. Copyright © Google LLC.

tensorflow.org
Google ML Kit Terms Apply

Detects faces and identifies objects in the camera view so Tolliver knows who and what he sees. Use is governed by Google APIs Terms of Service.

developers.google.com/ml-kit
Android Platform

Foundation & Libraries

CameraX Apache 2.0

Manages Tolliver's camera — the eyes he uses to see your family and surroundings. Copyright © The Android Open Source Project.

developer.android.com/training/camerax
OkHttp Apache 2.0

Handles Tolliver's internet communication — weather lookups and optional AI requests when you're online. Copyright © Square, Inc.

square.github.io/okhttp
Room (AndroidX) Apache 2.0

Stores Tolliver's memory — everything he's learned about your family — safely on your device. Copyright © The Android Open Source Project.

developer.android.com/training/data-storage/room
AndroidX & Material Apache 2.0

The foundation Tolliver's interface is built on — layouts, settings screens, and visual controls. Copyright © The Android Open Source Project.

developer.android.com/jetpack
Kotlin Apache 2.0

The programming language Tolliver is written in. Copyright © JetBrains s.r.o.

kotlinlang.org
Data & Services

Weather & Information

National Weather Service Public Domain

Provides Tolliver's weather data — current conditions, tonight's forecast, 7-day outlook. U.S. locations only. Free, no tracking. Published by the U.S. Government.

weather.gov
Optional Connected Services

Cloud Assistance

Google Gemini API Terms Apply

Tolliver may optionally connect to Google Gemini when a user provides their own API key. Gemini is never bundled with Tolliver. Each user connects on their own account. Use is subject to Google's API Terms and Policies.

ai.google.dev/gemini-api/terms
Before Public Release — License Review Required

Before Tolliver is published to the Google Play Store, any optional upgrade models — including future Llama-based models — should be reviewed carefully for commercial use restrictions and redistribution terms. Meta's Llama licensing has changed between versions and must be verified for the specific model used. This review is already planned as part of the pre-release checklist.

Tolliver stands on the shoulders of a global open-source community. These projects represent years of research, engineering, and generosity freely shared. We are grateful for every one of them.

Back to Tolliver