v0.7.2

Metadata patch release.

• Added explicit pub.dev platform metadata for Android, iOS, Linux, macOS, web, and Windows so the package listing reflects actual cross-platform support.
• Hardened the native prompt reuse CI model download to use authenticated, failing Hugging Face requests.

v0.7.1

• Apple native runtime packaging:
  • Added Flutter iOS/macOS Swift Package Manager integration so Apple apps link
    the pinned leehack/llamadart-native and leehack/litert-lm-native
    XCFramework artifacts through darwin/llamadart/Package.swift.
  • Disabled the legacy hook-managed Apple bundle path for Flutter iOS/macOS
    builds, avoiding wrapper/framework MinimumOSVersion mismatches in App
    Store uploads. Standalone Dart macOS keeps the native-assets dylib fallback.
  • Raised the Flutter Apple runtime floors to iOS 16.4 and macOS 14.0 to match
    the published XCFramework artifacts.
• Runtime defaults and release automation:
  • Android native builds still include both llama_cpp and litert_lm by
    default; iOS, macOS, Linux, and Windows now default to llama_cpp only.
  • Added native release pin automation so the maintainer sync workflow updates
    Apple SPM checksums from published native release asset digests.
• CI reliability:
  • Cached and retried tiny GGUF test-model downloads used by VM integration
    tests so main-branch CI is less exposed to Hugging Face 429 rate limits.
  • Excluded local SwiftPM artifact caches from pub archives; Flutter Apple
    consumers still resolve the published remote XCFramework targets.
• Compatibility note: no Dart API breaking changes. Flutter Apple apps must
  target iOS 16.4/macOS 14.0 or newer, and non-Android native apps that ship
  .litertlm models should opt in with llamadart_native_runtimes.

v0.7.0

• LiteRT-LM backend and runtime selection:
  • Added first-class .litertlm routing through LlamaBackend() on native and
    web targets, with native bundle downloads from leehack/litert-lm-native
    and web loading through @litert-lm/core.
  • Added ModelParams.liteRtLmBackend so callers can select LiteRT-LM CPU,
    GPU, or Android NPU execution where supported. auto chooses GPU on
    Android/macOS and CPU elsewhere on native targets.
  • Added cached Hugging Face .litertlm loading through
    loadModelSource(...), preserving the selected LiteRT-LM backend after the
    cache manager resolves the local file.
  • Added native LiteRT-LM tokenization, detokenization, log-level control,
    runtime metrics, and high-level ChatSession token counting support.
  • Added hooks.user_defines.llamadart.llamadart_native_tag,
    llamadart_native_repository, and llamadart_native_path so apps can test
    a different compatible native runtime source without patching llamadart.
  • Updated the Windows runtime fallback scanner to discover custom GitHub and
    local archive cache namespaces when .dart_tool/lib is unavailable.
• LiteRT-LM chat, templates, and generation quality:
  • Added GenerationParams.speculativeDecoding for native LiteRT-LM. The
    default remains disabled; llama.cpp, WebGPU, and LiteRT-LM web reject the
    option until their speculative paths are implemented.
  • Fixed Gemma 4 .litertlm thinking and tool calling by replacing the stub
    template with the canonical Gemma 4 chat template, parsing the runtime
    thought channel as reasoning, and suppressing reasoning deltas when callers
    set enableThinking: false.
  • Added a filename-keyed .litertlm chat-template registry seeded with
    Gemma 4/3/3n and Qwen 2.5/3. Pass ModelParams.chatTemplate to override
    detection for other models.
  • Fixed LiteRT-LM tool calling for grammar-using handlers by forwarding
    supportsGrammarConstraints from the active NativeAutoBackend delegate.
  • Stopped structured tool-call streams from leaking raw Hermes/Qwen JSON or
    Gemma <|tool_call> markers as assistant content before the final
    tool_calls chunk.
• Web and chat-app support:
  • Added ModelParams.preferMemory64 and ModelParams.modelBytesHint so
    large WebGPU GGUF models such as Gemma 4 E2B can choose the 64-bit bridge
    core before hitting the wasm32 address-space limit.
  • Fixed web .litertlm chat-app turns by swallowing unsupported token-count
    refreshes, avoiding unsupported minP/penalty parameters for LiteRT-LM
    web generation, and replacing the stuck "Loading model 0%" label with an
    indeterminate load message.
  • Halved web .litertlm load time by skipping WebGPU CacheStorage prefetch
    for LiteRT-LM models, which are fetched directly by @litert-lm/core.
  • Fixed web GGUF downloads reporting success before the bridge was ready by
    awaiting window.__llamadartBridgeReadyPromise, requiring the bridge
    prefetch API, and surfacing actionable errors for old bridge assets.
  • Allowed benign Hugging Face ?download=true URLs to be prefetched into the
    browser cache while still skipping credentialed or signed URLs.
• Lifecycle, cancellation, and native stability:
  • Fixed iOS .litertlm loading by resolving embedded LiteRtLm and
    StreamProxy frameworks from the app bundle, matching the macOS runtime
    path behavior.
  • Improved LiteRT-LM diagnostics before model load, including selected
    CPU/GPU/NPU backend reporting, platform availability errors, and complete
    dynamic-library candidate failures.
  • Validated platform-specific LiteRT-LM companion libraries during
    native-asset setup so incomplete runtime bundles fail at build time.
  • Hardened native and LiteRT-LM cancellation/disposal so in-flight generation
    no longer races token release, worker teardown, engine deletion, closed
    response ports, or stream writes after cancellation.
  • Freed multimodal prompt buffers on tokenize/eval error paths, serialized
    multimodal projector load/unload, and closed a leaked native-backend
    handshake reply port.
• Correctness and download resilience:
  • ChatSession now forwards empty-choices completion chunks instead of
    throwing, strips multiple <think> blocks, and trims history only on
    user-message turn boundaries.
  • LlamaEngine.generate wraps unexpected backend errors in
    LlamaInferenceException so callers catching LlamaException see the
    documented error type.
  • Tool-call parsing now uses stable fallback ids, tolerates code-fence
    language tokens without trailing delimiters, and keeps commas inside quoted
    argument values.
  • JSON-schema-to-GBNF conversion now resolves $refs nested inside other
    $ref targets and fails loudly on unresolvable or external $refs.
  • Array grammar generation validates minItems/maxItems, model downloads
    use connection and idle-read timeouts, and partial-download resume is
    restricted to files with stored validators.
• Benchmarks, docs, and validation:
  • Added fair Gemma 4 LiteRT-LM versus llama.cpp/GGUF benchmark tooling for
    Android, macOS, and web, with speculative-decoding metrics, Pixel benchmark
    failure detection, and target-specific timeouts.
  • Added tool/gguf_chat_features_smoke.dart and the
    chat-app-web-gemma4-webgpu-smoke E2E scenario for real-model parser and
    WebGPU mem64 validation.
  • Updated README, website docs, and doc/litert_lm_templates.md for backend
    selection, platform/runtime support, package-size controls, benchmark
    results, model templates, and current LiteRT-LM capability limits.
• Compatibility note: no public API breaking changes for existing GGUF /
  llama.cpp callers. LiteRT-LM support is additive, with deprecated benchmark
  wrappers retained for compatibility; unsupported llama.cpp-only parameters are
  rejected for .litertlm loads instead of being silently ignored.

v0.6.17

• Native runtime sync:
  • Updated native hook pinning and regenerated bindings through
    leehack/llamadart-native@b9371, picking up llama.cpp b9371.
  • Picked up the Apple mobile Metal stability fix that disables Metal
    residency sets on iOS/tvOS/visionOS native bundles, avoiding affected
    device context-creation failures such as MTLLibraryErrorDomain Code=3.
• Compatibility note: no public API breaking changes in 0.6.17;
  existing 0.6.16 callers remain compatible. The release only refreshes
  the pinned native runtime and generated low-level bindings.

v0.6.16

• Native runtime diagnostics:
  • Fixed native getVramInfo() so it reports free/total VRAM from
    llama.cpp GPU-class backend devices when available, using props-based
    memory reporting first and the legacy memory probe as a fallback.
  • Routed native VRAM probing through the ggml registry fallback path so
    Windows split bundles resolve backend-device symbols from the runtime that
    owns the device registry.
• WebGPU and chat app fixes:
  • Improved browser recovery for large remote WebGPU model/projector loads by
    retrying wasm32 model-staging aborts with the wasm64 core before surfacing
    memory-pressure failures.
  • Improved the runnable chat app's web remote-model startup path so model
    assets are prefetched into browser cache when available, browser
    CacheStorage failures fall back to direct network loading, and
    credentialed/signed model URLs skip persistent browser cache storage.
• Model download UX:
  • Improved the runnable chat app's mobile download behavior so lifecycle
    pauses no longer deliberately cancel active foreground downloads; the app
    now lets short screen-lock/background interruptions continue when the OS
    permits and still keeps explicit pause/dispose cancellation paths.
  • Added in-app and docs guidance for mobile large-model downloads, including
    resumable partial files, foreground Dart lifecycle limits, and the need for
    opt-in native background download/model-store integrations for robust
    cross-app GGUF management.
• Compatibility note: no public API breaking changes in 0.6.16;
  existing 0.6.15 callers remain compatible. The changes improve native VRAM
  diagnostics, WebGPU browser recovery, and chat app download lifecycle
  behavior.

v0.6.15

What's Changed

• chore(deps): bump path-to-regexp from 0.1.12 to 0.1.13 in /website in the npm_and_yarn group across 1 directory by @dependabot[bot] in #100
• chore(website): refresh docs dependencies by @leehack in #154
• test: add local E2E scenario runner by @leehack in #155
• Fix GLM-OCR multimodal prompt rendering by @leehack in #157
• Refactor chat template render context serialization by @leehack in #158
• Prepare llamadart 0.6.15 release by @leehack in #159

Full Changelog: v0.6.14...v0.6.15

v0.6.14

What's Changed

• docs: document WebGPU readiness checks by @leehack in #148
• chore(native): sync native release b9159 by @github-actions[bot] in #150
• feat: add model download controller by @leehack in #149
• ci: update actions for Node 24 by @leehack in #151
• chore: update WebGPU bridge assets to v0.1.16 by @leehack in #152
• chore: prepare v0.6.14 release by @leehack in #153

Full Changelog: v0.6.13...v0.6.14

v0.6.13

What's Changed

• fix: cascade WebGPU batch defaults by @leehack in #121
• feat: add model download cache manager by @leehack in #129
• feat(api): expose state_save_file / state_load_file on LlamaEngine by @thereisnotime in #123
• fix: serialize same-key model downloads by @leehack in #139
• fix: recover cache metadata sidecars by @leehack in #140
• docs: codify agent quality guidelines by @leehack in #142
• feat(webgpu): wire bridge state persistence by @leehack in #141
• fix: tolerate missing optional webgpu checksum assets by @leehack in #143
• docs: add production-readiness PR checklist by @leehack in #144
• fix(models): reject remote-only options for local sources by @leehack in #145
• feat(models): improve Hugging Face source ergonomics by @leehack in #146
• chore: prepare v0.6.13 release by @leehack in #147

Full Changelog: v0.6.12...v0.6.13

v0.6.12

What's Changed

• fix(native): test b9016 runtime and expose mainGpu by @leehack in #113
• Filter backend-owned runtime DLLs by @leehack in #115
• feat(api): expose load-time tuning knobs on ModelParams by @thereisnotime in #116
• Sync WebGPU bridge assets to llama.cpp b9016 by @leehack in #117
• Update WebGPU bridge assets to v0.1.14 by @leehack in #118
• Prepare 0.6.12 release by @leehack in #119

New Contributors

• @thereisnotime made their first contribution in #116

Full Changelog: v0.6.11...v0.6.12

v0.6.11

What's Changed

• chore(native): sync native release b8778 by @github-actions[bot] in #106
• fix(gemma4): stream thought channels as thinking by @leehack in #108
• chore(native): sync native release b8955 by @github-actions[bot] in #109
• chore(release): prepare 0.6.11 by @leehack in #110

Full Changelog: v0.6.10...v0.6.11