v1.3.1

What's Changed

• Fix duplicate detection for text-bearing insert_content_list calls by deferring doc_status creation until after LightRAG ainsert runs.
• Apply the same deferral to process_document_complete for parsed documents with text.
• Keep early doc_status creation for multimodal-only content that does not call ainsert.
• Add regression coverage for content-list and parsed-document duplicate handling.

Validation

• PYTHONPATH=. .venv/bin/python -m pytest tests/test_insert_content_list.py tests/testparser_wiring.py tests/testparser_kwargs.py -q
• uvx ruff check raganything/processor.py tests/test_insert_content_list.py raganything/__init__.py
• .venv/bin/python -m compileall raganything/processor.py tests/test_insert_content_list.py raganything/__init__.py

v1.3.0

What's Changed

⚠️ Behavior changes worth noting

• DoclingParser now uses the Docling Python API instead of shelling out to the docling CLI. This means:
  • You now need pip install docling to use it (the docling executable on PATH alone is no longer sufficient).
  • The env={...} kwarg on DoclingParser parse methods is still accepted for compatibility but is now ignored — set the relevant environment variables in the parent process or pass _get_converter kwargs (artifacts_path, table_mode, …).
  • <file_stem>.json and <file_stem>.md artifacts written under <output_dir>/<file_stem>/docling/ are still produced, but via export_to_dict() / export_to_markdown() rather than the CLI serializer — the logical content is the same but the files are not byte-identical.
  • check_installation() now tests Python importability rather than probing the CLI on PATH.
• MineruParser subprocess calls now run with a default timeout (configurable) and raise TimeoutError instead of hanging indefinitely.

✨ New features

• feat(parser): add remote URL support for DoclingParser by @bueno12223 in #195
• feat(omml): add OMML equation extraction utility for DOCX documents (closes #259) by @Abdeltoto in #262
• feat: add MiniMax provider support by @octo-patch in #264
• feat(examples): make LLM and vision model names configurable via env vars by @zhangzhenfei in #231
• feat: add Ollama integration example (closes #118) by @jwchmodx in #238

🛠 Refactor / performance

• refactor(parser): replace Docling CLI subprocess with Python API (closes #222) by @Abdeltoto in #261

🐛 Bug fixes

• fix: create doc_status even when LightRAG lacks multimodal insert args (closes #244) by @DeepaliPaspule in #255
• fix: prevent crashes from uninitialized LightRAG, env-var stripping, and parser cleanup by @jwchmodx in #240
• fix: add timeout parameter to MinerU subprocess to prevent indefinite hang (#172) by @peterCheng123321 in #254
• fix: pass entity_chunks_storage and relation_chunks_storage to all merge_nodes_and_edges calls (#241) by @peterCheng123321 in #250
• fix: handle messages= kwarg in vision_model_func (insert_content_list_example) (#28) by @peterCheng123321 in #252
• fix: forward system_prompt parameter in aquery_with_multimodal (#257) by @kuishou68 in #258
• fix(examples): preserve embedding kwargs with partial by @txhno in #263
• fix: demote misleading LibreOffice 'not found' warning to debug (closes #230) by @jwchmodx in #237
• fix: strip <think> tags from modal processor fallback responses (closes #159) by @jwchmodx in #236
• fix: create example log directory correctly by @haosenwang1018 in #242
• fix(init): remove duplicate __all__ assignment (#267) by @kuishou68 in #268
• fix: improve PDF parser handling by @davidangularme in #243

New Contributors

• @jwchmodx made their first contribution in #237
• @davidangularme made their first contribution in #243
• @zhangzhenfei made their first contribution in #231
• @bueno12223 made their first contribution in #195
• @kuishou68 made their first contribution in #268
• @peterCheng123321 made their first contribution in #250
• @txhno made their first contribution in #263
• @octo-patch made their first contribution in #264
• @Abdeltoto made their first contribution in #261
• @DeepaliPaspule made their first contribution in #255

Full Changelog: v1.2.10...v1.3.0

v1.2.10

What's Changed

• fix: use a single docling command for json and md formats by @wkpark in #198
• fix: normalize MinerU 2.0 field names for backward compatibility (#89) by @teamauresta in #202
• feat: add vLLM backend integration by @teamauresta in #201
• Fix potential path traversal and local file read vulnerabilities by @RinZ27 in #197
• feat(parser): add optional PaddleOCR backend by @SaqlainXoas in #199
• fix: prevent same-name file collision in parser output directories (#51) by @teamauresta in #203
• feat: add get_version() helper by @haosenwang1018 in #214
• feat: support environment variables in parsers by @wkpark in #210
• chore: export new public APIs in init.py by @Jah-yee in #219
• test: expand coverage for core config, utils, and batch parser by @Jah-yee in #218
• feat: add custom parser plugin system (closes #151) by @Jah-yee in #215
• feat: add processing events and callbacks system by @Jah-yee in #217
• fix(examples): use openai_embed.func to prevent double EmbeddingFunc wrapping by @syshin0116 in #223
• fix: use valid CID font names for Chinese text rendering (fixes #24) by @Exploreunive in #226
• fix: preserve full_entities metadata when adding multimodal entities by @Exploreunive in #228
• fix: handle closed event loop in close() to eliminate atexit warning (fixes #135) by @Exploreunive in #225
• feat: add retry and circuit breaker utilities for LLM calls (mitigates #172) by @Jah-yee in #216
• feat: add multilingual prompt template support (closes #85) by @Jah-yee in #220

New Contributors

• @wkpark made their first contribution in #198
• @teamauresta made their first contribution in #202
• @RinZ27 made their first contribution in #197
• @SaqlainXoas made their first contribution in #199
• @haosenwang1018 made their first contribution in #214
• @Jah-yee made their first contribution in #219
• @syshin0116 made their first contribution in #223
• @Exploreunive made their first contribution in #226

Full Changelog: v1.2.9...v1.2.10

v1.2.9

What's Changed

• feat: RAG-Anything runs offline by @LaansDole in #122
• Fix status comparison case mismatch in processor.py by @yrangana in #142
• [Doc]: fixing typos in a couple of tiles by @didier-durand in #162
• Feat query add system prompt by @EightyOliveira in #166
• Enhance the process_folder_complete function by @ikmak in #165
• Allow configuring full-path in file_path fields in RAG by @hanlianlu in #171
• using parser logger instead of logging by @ikmak in #180
• fix: resolve multiple bugs in RAGAnything initialization and parsing by @majiayu000 in #182
• Feature: batch dry run by @adevol in #185
• Fix #186: MinerU 2.7.0+ caused filepath errors with new default backend by @hanlianlu in #188

New Contributors

• @yrangana made their first contribution in #142
• @didier-durand made their first contribution in #162
• @EightyOliveira made their first contribution in #166
• @ikmak made their first contribution in #165
• @hanlianlu made their first contribution in #171
• @majiayu000 made their first contribution in #182
• @adevol made their first contribution in #185

Full Changelog: v1.2.8...v1.2.9

v1.2.8

What's Changed

• Add RAGAnything processing to LightRAG's webui by @hzywhite in #97
• Add RAGAnything processing to LightRAG's webui by @hzywhite in #113
• fix: replace del with atexit to fix RAGAnything cleanup warning by @liz-in-tech in #106
• feat: Add support for Chinese characters in PDF generation by @hongdongjian in #103
• Feat: LM Studio integration example and uv implementation by @LaansDole in #99

New Contributors

• @hzywhite made their first contribution in #97
• @liz-in-tech made their first contribution in #106
• @hongdongjian made their first contribution in #103
• @LaansDole made their first contribution in #99

Full Changelog: v1.2.7...v1.2.8

v1.2.7

What's Changed

• VLM Enhanced Query by @LarFii in #86

Full Changelog: v1.2.6...v1.2.7

v1.2.6

What's Changed

• Update .gitignore to include AI-related files and directories by @BenjaminX in #68
• Add Batch Processing and Enhanced Markdown Features by @ShorthillsAI in #64

New Contributors

• @BenjaminX made their first contribution in #68
• @ShorthillsAI made their first contribution in #64

Full Changelog: v1.2.5...v1.2.6

v1.2.5

What's Changed

• Direct Content List Insertion by @LarFii in #62
• Comprehensive Optimization of Multimodal Chunk Processing in RAGAnything by @LarFii in #65

Full Changelog: v1.2.4...v1.2.5

v1.2.4

What's Changed

• Add Docling Parser by @LarFii in #58
• Added caching mechanism by @LarFii in #59

Full Changelog: v1.2.3...v1.2.4

v1.2.3

What's Changed

• fix: _read_output_files 函数无法准确找到 md 文件的路径问题 by @nssai001 in #47
• 更新mineru2.0部分参数配置 by @liseri in #53

Hot Fix

• Fix issue#31

New Contributors

• @nssai001 made their first contribution in #47
• @liseri made their first contribution in #53

Full Changelog: v1.2.2...v1.2.3