Every Dart sample in the docs, one file per sample. Each file is the
doc's code block byte for byte: change both together, then run
python3 -B tool/samples_index.py and python3 -B tool/check_docs.py
(it fails, naming the doc and line, on any difference).
A dash in Modes is a mode the task offers that has no sample yet.
Platforms come from the coverage matrix (tool/coverage/matrix.json).
The runnable object detection app is in
apps/object_detection.
| Task | Modes | Other samples | Web | Android | iOS | macOS | Linux | Windows |
|---|---|---|---|---|---|---|---|---|
| common | browser_overlay (web only), capabilities | |||||||
| face_detector | image, video, live_stream | rgb_pixels | CPU, GPU | CPU | CPU (GPU planned) | CPU | CPU | CPU |
| face_landmarker | image, video, live_stream | count_faces, expressions_image, expressions_live_stream, expressions_video, gpu, model_from_asset, transformation_matrixes | CPU, GPU | CPU, GPU | CPU (GPU planned) | CPU | CPU, GPU | CPU |
| gesture_recognizer | image, video, live_stream | CPU, GPU | CPU | CPU (GPU planned) | CPU | CPU | CPU | |
| hand_landmarker | image, video, live_stream | CPU, GPU | CPU | CPU (GPU planned) | CPU | CPU | CPU | |
| holistic_landmarker | image, video, live_stream | mask_and_blendshapes | CPU, GPU | CPU | CPU (GPU planned) | CPU | CPU | CPU |
| image_classifier | image, video, live_stream | region_of_interest | CPU, GPU | CPU, GPU | CPU (GPU planned) | CPU | CPU, GPU | CPU |
| image_embedder | image, video, live_stream | quantized | CPU, GPU | CPU, GPU | CPU (GPU planned) | CPU | CPU | CPU |
| image_segmenter | image, video, live_stream | confidence_masks | CPU, GPU | CPU | CPU (GPU planned) | CPU | CPU, GPU | CPU |
| interactive_segmenter | image | CPU, GPU | CPU | CPU | CPU | CPU | no | |
| object_detector | image, video, live_stream | guide/create_image, guide/create_live_stream, guide/create_video, guide/input_image, guide/input_live_stream, guide/input_video, guide/results, guide/run_image, guide/run_live_stream, guide/run_video | CPU, GPU | CPU, GPU | CPU (GPU planned) | CPU | CPU, GPU | CPU |
| pose_landmarker | image, video, live_stream | segmentation_masks | CPU, GPU | CPU | CPU (GPU planned) | CPU | CPU | CPU |
| Task | Modes | Other samples | Web | Android | iOS | macOS | Linux | Windows |
|---|---|---|---|---|---|---|---|---|
| audio_classifier | clip, stream | capabilities | CPU | CPU | CPU | CPU | CPU | CPU |
| audio_transcriber | clip, stream | prewarm | CPU, GPU | CPU | CPU (GPU planned) | CPU, GPU | CPU | CPU |
| text_to_speech | utterance, stream | CPU | CPU | CPU | CPU, GPU | CPU | CPU |
| Task | Modes | Other samples | Web | Android | iOS | macOS | Linux | Windows |
|---|---|---|---|---|---|---|---|---|
| common | capabilities | |||||||
| language_detector | detect | CPU | CPU | CPU | CPU | CPU | CPU | |
| text_classifier | classify, top_category | CPU | CPU | CPU | CPU | CPU | CPU | |
| text_embedder | embed | CPU | CPU | CPU | CPU | CPU | CPU | |
| text_embedder (embedding_gemma model) | question_answering, semantic_similarity | CPU | CPU | CPU | CPU | CPU | CPU | |
| text_proofreader | proofread | no | CPU | CPU | CPU | CPU | CPU | |
| text_summarizer | capabilities, key_points, summarize | no | CPU | CPU | CPU | CPU | CPU |
| Task | Modes | Other samples | Web | Android | iOS | macOS | Linux | Windows |
|---|---|---|---|---|---|---|---|---|
| semantic_retriever | text, image, audio_and_content | manage_records, persistent_database | no | CPU | CPU | CPU | CPU | CPU |
| universal_embedder | text: -, image, audio | no | CPU | CPU | CPU | CPU | CPU |
| Task | Modes | Other samples | Web | Android | iOS | macOS | Linux | Windows |
|---|---|---|---|---|---|---|---|---|
| decision_maker | single, batch | no | CPU | CPU | CPU | CPU | CPU |
| Task | Modes | Other samples | Web | Android | iOS | macOS | Linux | Windows |
|---|---|---|---|---|---|---|---|---|
| model_store | clear, prefetch | |||||||
| web_runtime | self_hosted |