Basic app example

Path: example/basic_app · Platforms: Dart console on macOS, Linux and Windows · First-run download: Qwen3.5-0.8B-Q4_K_M.gguf (533 MB)

Dart console apps that show the core API: chat, embeddings, vector retrieval, decision models and image generation, without Flutter.

Run#

cd example/basic_app
dart pub get
dart run

Variants:

# Embeddings (downloads embeddinggemma-300M-Q8_0.gguf, 334 MB)
dart run bin/llamadart_embedding_example.dart -i "hello world" -i "rag"

# Retrieval: embeddings stored and searched in SQLite with sqlite_vector
dart run bin/llamadart_sqlite_vector_example.dart \
  -q "How do I improve embedding throughput?" \
  -d "Increase maxParallelSequences for wider embedding batches." \
  -d "Tune batchSize and ubatchSize together."

# Decision model: triage a support ticket
# (downloads laya-Q8_0.gguf, 421 MB, and laya-head.safetensors, 106 MB)
dart run bin/llamadart_decision_example.dart

# Image generation: write a PNG with SDXS (downloads the stable_diffusion
# runtime and the pinned sdxs-512-tinySDdistilled_Q8_0.gguf, 683 MB);
# Ctrl-C cancels
dart run bin/llamadart_image_example.dart -p "a red fox in autumn leaves"

What it demonstrates#

• Loading a model from an hf:// source into the shared cache, streaming a chat reply, and disposing the engine (Generation and streaming).
• LoRA adapters (--lora), GBNF-constrained output (--grammar) and a sample tool call (--tool-test) in the chat CLI (LoRA adapters, Tool calling).
• Single and batched embeddings, and query-versus-candidate probes (Embeddings).
• Local retrieval with SQLite vector search, exact or quantized, with a recall check against exact search (Embeddings).
• Typed choice, score and yes/no answers from a decision model with ChoiceKey.enumOf, ScoreKey.of, NoulKey.of and answerOf (Decision models).
• Experimental image generation with SDXS, SD-Turbo, or the desktop models (SDXL-Lightning, FLUX.1-schnell, SD 3.5 Large Turbo, Z-Image-Turbo) from the example's pinned downloads and settings: file and phase progress, Ctrl-C cancellation and PNG output, with the stable_diffusion runtime opted in through the example's pubspec.yaml (Image generation).

Test#

cd example/basic_app
dart test

Full options: every flag of the five CLIs, the retrieval result fields and the decision-model head options are in the example README.