docs: update docs for stable release - Production/Stable status, complete CHANGELOG, accurate README and AGENTS.md

Author: n24q02m

AGENTS.md

@@ -33,7 +33,8 @@ mise run fix       # ruff check --fix + ruff format
 ### Pytest Configuration
 
 - `testpaths = ["tests"]`, `pythonpath = ["."]`
-- Integration marker: `@pytest.mark.integration` (requires ~1.2 GB model download)
+- Integration marker: `@pytest.mark.integration` (requires model downloads: ONNX ~1.2 GB, Q4F16 ~1 GB, GGUF ~756 MB)
+- Integration test files: `test_integration.py` (ONNX), `test_integration_q4f16.py`, `test_integration_gguf.py`
 - CI runs: `uv run pytest -m "not integration" --tb=short`
 
 ## Code Style
@@ -53,7 +54,10 @@ mise run fix       # ruff check --fix + ruff format
 
 ### Type Checker (ty)
 
-Uses defaults (no custom config in pyproject.toml).
+Custom rules in `[tool.ty.rules]` (pyproject.toml):
+- `unresolved-import = "warn"` (llama_cpp is optional)
+- `possibly-missing-attribute = "warn"` (onnxruntime incomplete stubs)
+- `invalid-argument-type = "warn"`, `unresolved-attribute = "warn"`, `not-subscriptable = "warn"`, `invalid-assignment = "warn"`, `call-non-callable = "warn"`
 
 ### Import Ordering (isort via Ruff)
 

CHANGELOG.md

@@ -1,57 +1,81 @@
 # Changelog
 
-## [0.2.1](https://github.com/n24q02m/qwen3-embed/compare/v0.2.0...v0.2.1) (2026-02-14)
-
+## [1.2.0](https://github.com/n24q02m/qwen3-embed/compare/v1.1.3...v1.2.0) (2026-03-04)
 
 ### Features
 
-* promote dev to main (v0.2.1-beta) ([#5](https://github.com/n24q02m/qwen3-embed/issues/5)) ([8d8dbc1](https://github.com/n24q02m/qwen3-embed/commit/8d8dbc12ce60aec6b68b3c85ca5146b43de18f80))
+* migrate to 2025-2026 tech stack (uv/ty) ([7b689e4](https://github.com/n24q02m/qwen3-embed/commit/7b689e4))
+* add Codecov coverage upload ([891201b](https://github.com/n24q02m/qwen3-embed/commit/891201b))
+* add Renovate for automated dependency updates ([4e307b4](https://github.com/n24q02m/qwen3-embed/commit/4e307b4))
+* add comprehensive test coverage for utilities and components ([f1d2997](https://github.com/n24q02m/qwen3-embed/commit/f1d2997))
 
-## [0.2.1-beta](https://github.com/n24q02m/qwen3-embed/compare/v0.2.0...v0.2.1-beta) (2026-02-14)
+### Bug Fixes
 
+* make TextCrossEncoder.add_custom_model case-insensitive ([#88](https://github.com/n24q02m/qwen3-embed/issues/88)) ([30212f0](https://github.com/n24q02m/qwen3-embed/commit/30212f0))
+* fix path traversal and DoS risk ([#118](https://github.com/n24q02m/qwen3-embed/issues/118)) ([1bf8a49](https://github.com/n24q02m/qwen3-embed/commit/1bf8a49))
+* increase test coverage to 92% and resolve typing errors ([7658adf](https://github.com/n24q02m/qwen3-embed/commit/7658adf))
 
-### Features
+## [1.1.3](https://github.com/n24q02m/qwen3-embed/compare/v1.1.2...v1.1.3) (2026-02-28)
+
+### Bug Fixes
+
+* correct model identifiers from Qwen/ to n24q02m/ namespace ([298405c](https://github.com/n24q02m/qwen3-embed/commit/298405c))
 
-* add Q4F16 ONNX and GGUF model variant support ([8cc81c0](https://github.com/n24q02m/qwen3-embed/commit/8cc81c0ed10a3968758ba252dbc948aedc16cab1))
+## [1.1.2](https://github.com/n24q02m/qwen3-embed/compare/v1.1.1...v1.1.2) (2026-02-27)
+
+### Bug Fixes
+
+* GGUF Reranker Device.AUTO defaulting to CPU instead of GPU ([684af12](https://github.com/n24q02m/qwen3-embed/commit/684af12))
+
+## [1.1.1](https://github.com/n24q02m/qwen3-embed/compare/v1.1.0...v1.1.1) (2026-02-25)
+
+### Bug Fixes
 
+* auto-detect GPU for GGUF backend (Device.AUTO uses n_gpu=-1) ([edb58bb](https://github.com/n24q02m/qwen3-embed/commit/edb58bb))
 
 ### Documentation
 
-* update README with Q4F16 and GGUF variants ([c17f975](https://github.com/n24q02m/qwen3-embed/commit/c17f97541d579bda276b7138ed650deb1e7561d3))
+* add GPU acceleration section and fix CPU-only claim ([add0204](https://github.com/n24q02m/qwen3-embed/commit/add0204))
 
-## [0.2.0](https://github.com/n24q02m/qwen3-embed/compare/v0.1.0...v0.2.0) (2026-02-13)
+## [1.1.0](https://github.com/n24q02m/qwen3-embed/compare/v1.0.0...v1.1.0) (2026-02-22)
 
+### Features
 
-### ⚠ BREAKING CHANGES
+* auto-detect DirectML GPU provider and improve logging ([f400012](https://github.com/n24q02m/qwen3-embed/commit/f400012))
+
+## [1.0.0](https://github.com/n24q02m/qwen3-embed/compare/v0.2.1...v1.0.0) (2026-02-18)
+
+### Chores
+
+* migrate from release-please to python-semantic-release v10 ([9a78411](https://github.com/n24q02m/qwen3-embed/commit/9a78411))
 
-* Package import changed from 'from fastembed import ...' to 'from qwen3_embed import ...'. The PyPI package name remains qwen3-embed.
+## [0.2.1](https://github.com/n24q02m/qwen3-embed/compare/v0.2.0...v0.2.1) (2026-02-14)
 
 ### Features
 
-* promote dev to main (v0.2.0-beta) ([#2](https://github.com/n24q02m/qwen3-embed/issues/2)) ([4c95738](https://github.com/n24q02m/qwen3-embed/commit/4c95738222df7ae0ea339dc07d2a56d973e0365b))
-* qwen3-embed v0.1.0 - trimmed fastembed fork for Qwen3 models only ([d2a53fc](https://github.com/n24q02m/qwen3-embed/commit/d2a53fce71b55c44071717bd6722bcaf6fbcef7a))
+* add Q4F16 ONNX and GGUF model variant support ([8cc81c0](https://github.com/n24q02m/qwen3-embed/commit/8cc81c0))
+
+### Documentation
 
-## [0.2.0-beta](https://github.com/n24q02m/qwen3-embed/compare/v0.1.0...v0.2.0-beta) (2026-02-13)
+* update README with Q4F16 and GGUF variants ([c17f975](https://github.com/n24q02m/qwen3-embed/commit/c17f975))
 
+## [0.2.0](https://github.com/n24q02m/qwen3-embed/compare/v0.1.0...v0.2.0) (2026-02-13)
 
 ### ⚠ BREAKING CHANGES
 
-* Package import changed from 'from fastembed import ...' to 'from qwen3_embed import ...'. The PyPI package name remains qwen3-embed.
+* Package import changed from `from fastembed import ...` to `from qwen3_embed import ...`. The PyPI package name remains `qwen3-embed`.
 
 ### Features
 
-* qwen3-embed v0.1.0 - trimmed fastembed fork for Qwen3 models only ([d2a53fc](https://github.com/n24q02m/qwen3-embed/commit/d2a53fce71b55c44071717bd6722bcaf6fbcef7a))
-
+* qwen3-embed v0.1.0 - trimmed fastembed fork for Qwen3 models only ([d2a53fc](https://github.com/n24q02m/qwen3-embed/commit/d2a53fc))
 
 ### Bug Fixes
 
-* port ONNX compatibility fixes and add integration tests ([59e30d0](https://github.com/n24q02m/qwen3-embed/commit/59e30d00138668921763592d357ea68efb0decc1))
-* resolve CI lint failures and exclude integration tests from CI ([a645d52](https://github.com/n24q02m/qwen3-embed/commit/a645d524a3f0b322e62fad3baf758995c19decc9))
-
+* port ONNX compatibility fixes and add integration tests ([59e30d0](https://github.com/n24q02m/qwen3-embed/commit/59e30d0))
 
 ### Code Refactoring
 
-* rename package from fastembed to qwen3_embed ([6e77ba5](https://github.com/n24q02m/qwen3-embed/commit/6e77ba5715d48e33fe8bb863a135e17b2a385a1a))
+* rename package from fastembed to qwen3_embed ([6e77ba5](https://github.com/n24q02m/qwen3-embed/commit/6e77ba5))
 
 ## 0.1.0 (2026-02-13)
 

README.md

@@ -1,6 +1,6 @@
 # Qwen3 Embed
 
-**Lightweight Qwen3 text embedding and reranking via ONNX Runtime**
+**Lightweight Qwen3 text embedding and reranking via ONNX Runtime and GGUF**
 
 [![CI](https://github.com/n24q02m/qwen3-embed/actions/workflows/ci.yml/badge.svg)](https://github.com/n24q02m/qwen3-embed/actions/workflows/ci.yml)
 [![codecov](https://codecov.io/gh/n24q02m/qwen3-embed/graph/badge.svg?token=M038M651L2)](https://codecov.io/gh/n24q02m/qwen3-embed)
@@ -148,7 +148,8 @@ model = TextEmbedding(model_name="n24q02m/Qwen3-Embedding-0.6B-ONNX", cuda=Devic
 
 ### GGUF
 
-GPU is handled by `llama-cpp-python`. Install with CUDA support:
+GPU is handled by `llama-cpp-python`. The default `pip install qwen3-embed[gguf]` is CPU-only.
+For CUDA GPU support, build with:
 
 ```bash
 CMAKE_ARGS="-DGGML_CUDA=on" pip install qwen3-embed[gguf]
@@ -175,7 +176,6 @@ mise run fix     # ruff auto-fix + format
 
 ## Related Projects
 
-- **[modalcom-ai-workers](https://github.com/n24q02m/modalcom-ai-workers)** — GPU-accelerated AI workers on Modal.com using this embedding model
 - **[mnemo-mcp](https://github.com/n24q02m/mnemo-mcp)** — MCP memory server with semantic search powered by qwen3-embed
 - **[wet-mcp](https://github.com/n24q02m/wet-mcp)** — MCP web search server with vector-based docs search
 

pyproject.toml

@@ -1,16 +1,17 @@
 [project]
 name = "qwen3-embed"
 version = "1.2.0"
-description = "Lightweight Qwen3 text embedding & reranking via ONNX Runtime (fork of fastembed)"
+description = "Lightweight Qwen3 text embedding & reranking via ONNX Runtime and GGUF (fork of fastembed)"
 readme = "README.md"
 license = "Apache-2.0"
 requires-python = ">=3.13"
 authors = [{ name = "n24q02m" }]
 keywords = ["vector", "embedding", "reranking", "qwen3", "onnx", "onnxruntime"]
 classifiers = [
-    "Development Status :: 3 - Alpha",
+    "Development Status :: 5 - Production/Stable",
     "Intended Audience :: Developers",
     "License :: OSI Approved :: Apache Software License",
+    "Programming Language :: Python :: 3",
     "Programming Language :: Python :: 3.13",
     "Topic :: Scientific/Engineering :: Artificial Intelligence",
 ]