infiniflow/ragflow v0.25.1 on GitHub

Summary

Improvements

API refactoring and unification: Standardizes web APIs to RESTful conventions across all endpoints, unifying document creation and indexing flows while maintaining backward compatibility.
Parsing optimizations: Adds OpenDataLoader PDF backend. #14097
Introduces lazy loading and chunked parsing for large PDFs (>50 pages), significantly reducing memory footprint. #14385

Data source

Enables synchronizing deleted files in Bitbucket, Gmail, Google Drive, and Airtable.

Model support

DeepSeek v4

Model providers

UCloud

Bug fixes

Metadata visibility issues during v0.24.0 to v0.25.0 upgrades.
Duplicate chat output.

What's Changed

Refactor: Consolidation WEB API & HTTP API for document get_filter by @xugangqiang in #14248
docs: add DeepWiki developer guide page by @hyl64 in #14244
Docs: User-level memory is supported in v0.25.0 by @writinwaters in #14259
Refactor: Consolidation WEB API & HTTP API for document infos by @xugangqiang in #14239
Go: add balance command by @JinHai-CN in #14262
Update release note by @JinHai-CN in #14275
Refa: align chat and search restful APIs by @buua436 in #14229
Refactor: Consolidation WEB API & HTTP API for document delete api by @xugangqiang in #14254
fix: normalize think tags in final chat answer by @buua436 in #14271
Fix: switch MinerU API endpoint to /pdf_parse by @6ba3i in #14272
Doc: v0.25.0 release notes. by @writinwaters in #14284
Fix #14213 create folder does not accept FOLDER by @wangq8 in #14276
Fix uv.lock by @JinHai-CN in #14285
Fix: document and sdk support of searching message with user_id by @Lynn-Inf in #14283
Fix: MinerU 3.x output discovery and API contract by @6ba3i in #14282
feat: Add Astraflow provider support (global + China endpoints) by @ucloudnb666 in #14270
Add deepseek and moonshot model json by @JinHai-CN in #14290
Fix upload stream handling to prevent truncated files by @bohdansolovie in #14267
Fix: serialize GraphRAG entity resolution merges to avoid graph mutation races by @spider-yamet in #14237
Set image tag v0.25.0 by @wangq8 in #14299
Doc: two PDF parser optimizers are supported as of v0.25.0. by @writinwaters in #14261
Refact: Tenant api by @Magicbook1108 in #14288
Refactor: Migrate document metadata config update API by @xugangqiang in #14286
Refactor: API connectors by @wangq8 in #14228
Fix: Remove duplicate text output from the thought model on the chat page. by @cike8899 in #14301
Fix: Some bugs by @dcc123456 in #14287
Go: add new provider minimax by @JinHai-CN in #14296
Fix: Recall Test Page Metadata Not Displaying. by @cike8899 in #14297
Refactor: REST API langfuse api-key by @wangq8 in #14315
Refactor: API file2document by @wangq8 in #14306
Refactor: Doc metadata update by @xugangqiang in #14289
Refa: migrate MCP APIs to RESTful api by @buua436 in #14317
Add REDIS zcard by @wangq8 in #14316
Build(deps): Bump lxml from 6.0.2 to 6.1.0 in /sdk/python by @dependabot[bot] in #14318
Add missing timeout to ragflow server health check by @Ricardo-M-L in #14311
Refact: system apis by @Magicbook1108 in #14298
Refa: migrate chunk APIs to RESTful routes by @buua436 in #14291
API refactor: stats_api and plugin_api by @wangq8 in #14324
Fix commit override from #14298 of api-key to api_key by @wangq8 in #14328
Feat: optimize title chunk by @Magicbook1108 in #14325
Refa: remove legacy MCP server web API by @buua436 in #14322
fix azure blob put method param by @newyangyang in #14329
Feat: Agent api by @Magicbook1108 in #14157
fix: migrate Langfuse integration from start_generation to start_obse… by @RazmikGevorgyan in #14205
Refactor user REST API by @wangq8 in #14334
Feat: deepseek v4 by @Magicbook1108 in #14346
Implement retrieval_test in GO by @qinling0210 in #14231
docs: fix API key guide typo by @MukundaKatta in #14352
Fix api user patch verb does not work by @wangq8 in #14358
Refa : migrate agent webhook routes to REST APIs by @buua436 in #14330
Fix: allow use image2text as chat model by @Lynn-Inf in #14331
feat: route docling parsing through native chunking endpoints by @ParasSondhi in #14218
fix: check isinstance before len in VariableAssigner _remove_first/_remove_last by @kuishou68 in #14281
Fix blob sync: skip unsupported files before download by @6ba3i in #14357
Refact: Updated rootAsHeadingTip by @writinwaters in #14363
Fix: The button styles in the PaddleOCR dialog are not applying correctly. by @cike8899 in #14350
Update API document by @wangq8 in #14364
Doc: Updated a 0.25-specific faq by @writinwaters in #14365
Go: add gitee and siliconflow as model provider by @JinHai-CN in #14336
Feat: introduce minimum type check for pipeline by @Magicbook1108 in #14354
Fix: allow search id or _id by @Lynn-Inf in #14356
Feat: add OpenDataLoader PDF parser backend (#14058) by @wdeveloper16 in #14097
Fix: validate URL scheme and resolved IP before crawling to prevent SSRF by @xingxing21 in #14090
feat(api): add unified index API and dataset management endpoints by @euvre in #14222
Refa: unify document create flows under REST documents API by @buua436 in #14345
feat: persist RAPTOR layer metadata on summary chunks by @yuch85 in #13286
fix: skip canvas SSE fetch in chat shared page to eliminate spurious 103 error by @euvre in #14190
fix: handle Infinity table-not-exist error (3022) in update() methods by @euvre in #14153
feat: persist PDF bookmark outline as document metadata by @yuch85 in #13287
chore(CLAUDE.md): add shared UI component lock convention to CLAUDE.md by @ZhenhangTung in #14381
Refa: restore openai-compatible chat completions api by @buua436 in #14380
Go: aliyun model provider by @JinHai-CN in #14379
Fix: Remove hardcoded page limits causing parsing failures on large PDFs (>300 pages) by @euvre in #14382
Refactor: migrate artifact API by @xugangqiang in #14348
Remove evaluation_app.py and kb_app.py by @wangq8 in #14394
Fix metadata parsing regression for upgraded v0.24 datasets by @6ba3i in #14383
perf: lazy img_np loading and chunked parse_into_bboxes for large PDFs by @yuzhichang in #14385
Refactor: migrate doc upload info used in chat by @xugangqiang in #14359
Fix: support release in agent update api by @buua436 in #14396
Helm template ragflow.yaml: fix nginx-config-volume mountPath according to Dockerfile v0.25.0 by @mginfn in #14361
Fix: prioritize explore session ID and reset default conversation variables by @buua436 in #14399
Refactor: deco doc-parse API that is not used any more by @xugangqiang in #14367
Refa: align list operations and strict mode by @buua436 in #14387
Add task API by @wangq8 in #14393
Refactor: optimize agent reset conversation variable defaults by @buua436 in #14401
Refactor: Doc batch change status by @xugangqiang in #14337
tests: add missing HTTP API tests for dataset management endpoints removed in #14222 by @euvre in #14390
Refactor: deco document upload_and_parse API by @xugangqiang in #14366
Go: add new provider: google by @JinHai-CN in #14395
Refactor: migrate document run api by @xugangqiang in #14351
Refactor: migrate document thumbnails API by @xugangqiang in #14344
Fix: add executor.shutdown by @xugangqiang in #14403
Refactor: Doc change parser by @xugangqiang in #14327
Doc: Added a database schema and migration guide. by @writinwaters in #14404
Fix: thumbnails issue in chat by @xugangqiang in #14415
Go: add volcengine by @JinHai-CN in #14409
Fix: preserve infinity available_int zero filter by @buua436 in #14416
Fix: align chat recommendation and thumbup APIs by @buua436 in #14413
Always return success if no such task id by @wangq8 in #14417
Refactor model in GO by @qinling0210 in #14398
Go: fix compilation by @JinHai-CN in #14418
Fix manual naive parser position extraction fallback by @6ba3i in #14420
Fix case-insensitive matching for manual meta_data_filter in / not in list values by @spider-yamet in #14397
Feat: enable sync deleted files for more connectors by @Magicbook1108 in #14353
Go: update db model by @JinHai-CN in #14423
Refactor model type to model class by @JinHai-CN in #14426
Fix: agent toolcall null response & schema validation & DeepSeek think history by @buua436 in #14425
Fix: document level auto metadata config missing after save by @xugangqiang in #14421
Go: implement provider: Moonshot by @Haruko386 in #14433
Feat: more model for paddle by @Magicbook1108 in #14436
Simplify Encode by @qinling0210 in #14437
Fix: google authentication - gmail && google-drive by @Magicbook1108 in #14422
Refactor: migrate chunk retrieval_test and knowledge_graph to REST API endpoints by @euvre in #14402
Fix: update based on #14436 by @Magicbook1108 in #14440
Fix: enable sync deleted file in airtable by @Magicbook1108 in #14438
refactor: improve QwenRerank logic by @Woody-Hu in #14388
feat(google-drive): optimize memory payload and enable sync deletion by @ParasSondhi in #14372
Feat: sync deleted files in Bitbucket by @Magicbook1108 in #14450
Go: update chat URL by @JinHai-CN in #14453
Fix: prune deleted doc chunks from retrieval by @buua436 in #14454
Fix: Dataset: When configuring the "general chunk method," options such as chunk size and parent-child slicing are unavailable. by @cike8899 in #14459
Remove model_bundle.go, modify chat_session.go by @qinling0210 in #14458
Fix: add retrieval fallback comments by @buua436 in #14457
Add backward compat APIs by @wangq8 in #14427
Go: implement provider: volcengine by @Haruko386 in #14460
Fix delete graphrag raptor by @wangq8 in #14469
Fix query param type by @wangq8 in #14471
Fix: Clicking the button in the bottom-right corner of the /chats/widget page fails to display the dialog box. by @cike8899 in #14465
Feat: enable sync deleted files for Gmail && fix google drive issues by @Magicbook1108 in #14462
Port PR14454 to GO (PruneDeletedChunks) by @qinling0210 in #14463
Update create model instance command by @JinHai-CN in #14441
Fix graph task type by @wangq8 in #14475
Fix delete graph by @wangq8 in #14484
Fix: RAPTOR "Generation scope" reset to "Single file" when selecting "Dataset" by @euvre in #14477
Feat: enable sync deleted files in gitlab by @Magicbook1108 in #14481
feat(dropbox): support deleted-file sync by @bitloi in #14476
Feat: enable sync deleted file for Discord by @Magicbook1108 in #14451
Go: implement provider: MiniMax by @Haruko386 in #14478
Go: add drop instance models by @JinHai-CN in #14485
Docs: Updated Title chunker references by @writinwaters in #14483
Fix: The pipeline column header in the FileLogsTable is displaying incorrectly. by @cike8899 in #14489
Fix visit dataset error by @wangq8 in #14490
Docs: by @writinwaters in #14492
Fix metadata config by @wangq8 in #14480
fix: file logs not displayed in dataset ingestion page by @euvre in #14479
Chore(deps): Bump google.golang.org/grpc from 1.66.2 to 1.79.3 by @dependabot[bot] in #14513
Fix: The GraphRAG icon is not displaying. by @cike8899 in #14514
Docs: Update version references to v0.25.1 in READMEs and docs by @asiroliu in #14488

New Contributors

@hyl64 made their first contribution in #14244
@ucloudnb666 made their first contribution in #14270
@bohdansolovie made their first contribution in #14267
@newyangyang made their first contribution in #14329
@RazmikGevorgyan made their first contribution in #14205
@MukundaKatta made their first contribution in #14352
@ParasSondhi made their first contribution in #14218
@kuishou68 made their first contribution in #14281
@wdeveloper16 made their first contribution in #14097
@xingxing21 made their first contribution in #14090
@yuch85 made their first contribution in #13286
@mginfn made their first contribution in #14361
@Haruko386 made their first contribution in #14433

Full Changelog: v0.25.0...v0.25.1