google-ai-edge

Show an interactive map view for the given location.

A daily learning companion that teaches users a new concept, generates a beautiful image card, and schedules recurring daily learning notifications.

Schedule a notification for a specific date or repeating daily.

Read OS calendar events for a specific date.

Create a calendar event.

Calculate the hash of a given text.

Act as a dungeon master for a text-based adventure set in a world where everyone is a sentient kitchen appliance. Trigger when user says "start kitchen adventure".

A simple mood tracking skill that stores your daily mood and comments. Use this when the user wants to log their mood, track how they feel, or see their mood history.

Build a new Android app (Kotlin, Compose) around a verified LiteRT model using the CompiledModel API - the app architecture, the inference-layer lifecycle rules, model delivery, and the UI traps that masquerade as model bugs. Use when turning a converted and…

Convert a PyTorch or Hugging Face model into a LiteRT model that runs fully on the GPU via the CompiledModel API with verified-correct output, and lay it out as a model recipe. Use when converting a new model, or when a converted model is rejected by the GPU,…

Rapidly migrate an Android application from legacy TensorFlow Lite (TFLite) to the modern LiteRT CompiledModel API (2.x) in Open Source GitHub repositories. Covers buffer reuse, NPU JIT compilation, a reference implementation, and a two-stage verification…

Creates an Android app that runs a .tflite model on the CPU or the GPU with the LiteRT CompiledModel API in Kotlin. Use this skill to build a new app around a vision, audio or embedding model, or to add on-device inference to an existing app - the dependency,…

Shrink a converted LiteRT model with ai-edge-quantizer (fp16 / int8 / int4) without losing accuracy, verifying parity against the float source after every step. Use when choosing a quantization recipe for a new model, when a quantized model fails to load,…

Convert a Hugging Face LLM or vision-language model checkpoint into a .litertlm bundle that runs on the LiteRT-LM runtime with verified quality - classify the architecture against known runtime walls, pick the recipe family (dense, reasoning, hybrid SSM,…

Measure a LiteRT .tflite model with `benchmark_model` on Developer Device Platform (DDP) lab phones through `litert benchmark --ddp`, or on the Mac you run it from, and turn the session into rows of the performance leaderboard under benchmark/leaderboard -…

Prove a converted or quantized LiteRT model on the actual device via the CompiledModel API - confirm GPU residency, compare device output against the source model, and diagnose device-only failures such as silent CPU fallback, whole-graph compile ceilings,…

Diagnoses and fixes on-device GPU numerical corruption, garbage generation, zeroed KV caches, and ML Drift delegate lowering bugs for LiteRT and LiteRT-LM models on Android (OpenCL and WebGPU). Use when a model produces correct output on CPU but generates…

Authors, registers, and tests new operator test generators for the LiteRT Accelerator Test Suite (ATS) in litert/test/generators/ and litert/ats/. Use when onboarding a new TFLite primitive op or StableHLO composite op into LiteRT ATS, expanding ATS op…

Systematically explore quantization configurations for a TFLite model using the LiteRT Quantizer API, finding the optimal recipe that balances file size and accuracy. Use this skill whenever the user wants to quantize a model, optimize a recipe, explore…

Converts PyTorch models (e.g. ResNet, timm, HuggingFace transformers) directly to LiteRT (.tflite) flatbuffer format. Use when converting PyTorch models to TFLite, setting up export environments, or troubleshooting torch-to-litert conversion bugs. Don't use…

Validates equivalence between LiteRT models (litert_lm) and PyTorch models (transformers). Use when you need to verify that an exported LiteRT model produces the same outputs as the original Hugging Face model. Supports multi-turn conversations and custom…

Assists the user to calibrate, merge, and statically quantize litert LLM models (such as Gemma 3) in standard open-source (OSS) environments. Use when the user wants to run LLM calibration, merge task JSON results, align KV cache parameters across models,…

Guides the agent to orchestrate and implement a standalone LiteRT-LM Android demo app with backend selection and multi-modality support using Bazel 9. Use when asked to create a demo app for LiteRT-LM on Android. Don't use for general Android app development…

Systematically explore quantization configurations for a TFLite model using the AI Edge Quantizer API, finding the optimal recipe that balances file size and accuracy. Use this skill whenever the user wants to quantize a model, optimize a recipe, explore…

Generates code for a new GPU Kernel in ml_drift using UCL syntax. Use when the user wants to add or implement a new operation/kernel in ml_drift (e.g., SquaredDifference, Maximum). Creates the .h, .cc files, generates test files (test.cc, test_util.h,…

LiteRT CLI tool to download, convert, quantize, run, benchmark, and visualize LiteRT models.