infiniflow/ragflow v0.26.0

on GitHub

Summary

New features

• Model providers
  • Implements auto-populated model lists for multiple providers, eliminating the need to type model names manually. This feature currently supports: Ollama, OpenRouter, vLLM, OpenAI-API-Compatible, LM-Studio, VolcEngine, Xinference, LocalAI, BaiduYiyan, GPUStack, and Fish Audio.
  • Allows configuring multiple API keys for the same model provider. #14595
  • Dynamically populates model selection dropdowns in the UI by fetching the currently available models directly from remote model providers. #15711
• Data source connectors: Implements new data source connectors for Outlook, OneDrive, Microsoft Teams, Slack, SharePoint, Salesforce, and Azure Blob Storage. #15333#15330#15332#15188#15190#15462#15466
• Dataset - Implements a checkpoint and resume feature for community extraction and entity resolution, the most expensive and time-consuming parts of the GraphRAG indexing pipeline. #15518#15523

Improvements

• Removes <think> text buffering to ensure reasoning-capable models feel faster and more transparent during interactions. #15891
• Marks MySQL migrations as applied. #15504

Model Support

• Four new SiliconFlow models #15383
• MiniMax-M3 model #15513
• Latest Anthropic models #15516
• Voyage 4 model family #15516
• Cohere model list. #15576

i18n

• Completes Korean translation. #15863
• Completes Italian translation. #15729

Bug fixes

• The thinking mode of MiniMax models was not correctly enabled. #15496
• Infinite loops were triggered when the thinking mode was enabled for Qwen3.5 and Qwen3.6 models. #15101
• Streamed answers were being duplicated when using the OpenAI-compatible chat completions API endpoint. #15286#15443
• Serialization errors were caused during chat completions when invalid numeric scores like NaN (Not-a-Number) or Inf (Infinity) were passed to the JSON encoder. #15245#15266
• Chat completions using LiteLLM providers were failing because unrecognized internal configuration parameters were not being filtered out before reaching the external APIs. #15427#15432
• The OpenAI-compatible chat completions API was defaulting to streamed responses. #15356#15394
• Empty AND results were incorrectly dropped during metadata filtering. #15477
• Repetitive page chrome, such as headers and footers, was incorrectly extracted as main text by the MinerU parser. #15335#15387
• English chart titles were missing during document extraction in the DeepDoc module. #15481
• Empty outputs were returned by the TitleChunker for json and chunks upstream formats #14247#15396
• An error message was missing when a .tsv file upload attempt failed. #15284
• API tokens missing beta values caused token retrieval errors. #15405
• Caps the maximum page size to fix system crashes or slowdowns from large queries. #15292
• Client errors were caused by the OpenAI-compatible chat completion API incorrectly defaulting to streamed responses. #15356#15394
• HTTP 500 internal server errors were triggered instead of standard 4xx client errors when users attempted to download missing files from the storage backend. #15369#15371
• GraphRAG entity ranking was broken. #15795#15797

What's Changed

• Fix: The Creativity parameter of chat was not saved. by @cike8899 in #15243
• Feat: add new tests and tescases for restful api suite by @6ba3i in #15277
• fix(go-models): add VolcEngine model listing suffix by @oktofeesh1 in #15234
• go: add Qiniu model provider by @Hz-186 in #15280
• feat: add GiteeAI provider support to Go API server by @Helios531 in #15131
• Go: implement Rerank in Replicate driver by @sxxtony in #15278
• Go: implement connector get API by @tmimmanuel in #15259
• Go: implement TTS and ASR for X.AI by @Haruko386 in #15247
• fix(go-models): validate TokenHub chat requests by @oktofeesh1 in #15283
• Fix: Uploading TSV format documents to the knowledge base did not generate any error messages. by @cike8899 in #15284
• feat[Go]: implement create_connector API by @Haruko386 in #15285
• Fix: ExeSQL node continues on per-statement SQL errors by @nickmopen in #15140
• Docs: Updated converse with agent API reference. by @writinwaters in #15257
• Docs: Finalized v0.25.6 release notes. by @writinwaters in #15305
• Refactor: Task Executor by @xugangqiang in #15154
• go: add PATCH /api/v1/users/me user settings update by @Hz-186 in #15297
• Feat: add new tests and tescases for restful api suite by @6ba3i in #15299
• Feature: Allow page_size max value 100 by @wangq8 in #15292
• feat: implement SharePoint data source connector by @JPette1783 in #15190
• Go: implement system healthz API by @tmimmanuel in #15307
• fix: include user model settings in /user/me response by @Hz-186 in #15320
• fix(go-models): harden 302.AI driver requests by @oktofeesh1 in #15289
• Fix: Clustering method echo error by @dcc123456 in #15322
• feat: implement Slack data source connector by @JPette1783 in #15188
• Update developer doc by @xugangqiang in #15336
• feat: add Go MCP server create API by @Helios531 in #15260
• feat[Go]: implement delete/ rebuild/ listlog api for connector by @Haruko386 in #15300
• fix(go-models): harden LongCat default transport handling by @glorydavid03023 in #15340
• fix(go-models): harden Voyage default transport handling by @glorydavid03023 in #15341
• feat: implement Microsoft Teams data source connector by @JPette1783 in #15193
• feat(go-api): implement connector (data source) management endpoints by @JPette1783 in #15274
• feat(go-api): implement GET /api/v1/agents list endpoint (issue #15328) by @JPette1783 in #15329
• fix(go-models): harden ModelScope default transport handling by @glorydavid03023 in #15339
• Fix(kb): enforce tenant authorization on UpdateMetadataSetting (#15268) by @nickmopen in #15270
• fix(go-models): harden JieKouAI driver requests by @oktofeesh1 in #15337
• feat[Go]: implement /api/v1/system/status GET by @Haruko386 in #15348
• feat(go-api): implement tenant member management (issue #15294) by @JPette1783 in #15295
• fix: validate custom model inputs by @bitloi in #15200
• fix(go-models): harden Novita default transport handling by @glorydavid03023 in #15350
• Feat: tenant llm provider by @Lynn-Inf in #14595
• Fix: default OpenAI chat completions to non-stream by @buua436 in #15394
• feat(connector): implement OneDrive data source connector (issue #15330) by @JPette1783 in #15331
• feat: add 4 new models for siliconflow by @leslie2046 in #15383
• Fix user registration initialization in Go API by @Hz-186 in #15349
• feat: add Go MCP server delete API by @Helios531 in #15262
• fix: persist logout access token invalidation by @Hz-186 in #15397
• feat[Go]: implement /system/stats and refactor /system/config/log by @Haruko386 in #15407
• fix(go): resolve custom base_url for empty default region by @bitloi in #15043
• fix(metadata): preserve empty AND results across filter conditions by @monsterdavidliu-ux in #15386
• fix(go-models): raise SSE scanner buffer so large stream chunks are not dropped by @Dexterity104 in #15382
• Validate chunk image_base64 before doc-store write by @kiannidev in #15364
• Fix: JWT algorithm-confusion in OIDC ID token verification by @galuis116 in #15181
• fix(go-models): harden Hunyuan embedding validation by @oktofeesh1 in #15249
• fix: backfill missing api token beta values by @Hz-186 in #15405
• File addition .rooignore by @Attili-sys in #15414
• fix: upgrade crawl4ai to 0.8.0 (CVE-2026-26217) by @orbisai0security in #15415
• fix(api): move dify retrieval health check to /dify/retrieval/health by @jony376 in #15311
• fix(opensearch): implement get_scores for KNN second-pass scoring by @ef-rintaro in #15390
• fix(opensearch): repair document-metadata path broken by #14577 by @ef-rintaro in #15393
• feat(connector): implement Outlook data source connector (issue #15332) by @JPette1783 in #15333
• Fix: If the filename is too long, it overflows the confirmation box for deleting the file. by @cike8899 in #15287
• Fix memory empty issue by @wangq8 in #15411
• Feat: add new tests and tescases for restful api suite by @6ba3i in #15347
• feat[Go] implement api/v1/thumbnails API by @Haruko386 in #15416
• Go: implement Embed in GPUStack driver by @glorydavid03023 in #15182
• Go: implement dataset ingestion log APIs by @sxxtony in #15421
• GraphRAG - NER - spacy - fix spacy extraction by @wangq8 in #14783
• Fix: Unable to create dataset by @cike8899 in #15472
• fix: restore TitleChunker output for json/chunks upstream formats by @euvre in #15396
• Fix: The newly added model did not appear in the drop-down menu. by @cike8899 in #15476
• Return 4xx when file download storage blob is missing by @kiannidev in #15371
• Infer Content-Type for document image endpoint by @kiannidev in #15368
• Bug fix: Enhance embeding model to give better error message by @wangq8 in #15346
• fix(deepdoc): add English caption patterns to fix missing figure/table numbering by @euvre in #15481
• fix(go-models): apply custom Google base URLs by @oktofeesh1 in #15385
• fix(mineru): skip page chrome blocks to prevent duplicate chunks by @monsterdavidliu-ux in #15387
• Docs: Added a guide on how to ingest an RSS feed. by @writinwaters in #15467
• docs: add Tigris as S3-compatible storage backend, fix s3 region field name by @davidmyriel in #15361
• feat: add Go MCP server list API by @Helios531 in #15253
• feat(go-api): implement password-reset flow (issue #15282) by @JPette1783 in #15293
• Fix:empty chat model fallback by @buua436 in #15477
• fix(llm): strip non-generation keys from gen_conf for LiteLLM providers (#15427) by @nickmopen in #15432
• fix(chat): sanitize NaN/Inf scores before serializing chat completions (#15245) by @Rene0422 in #15266
• Fix auto metadata type issue by @xugangqiang in #15338
• fix(api): handle array message content on OpenAI chat completions by @glorysr1209-png in #15359
• fix(migration): parameterize tenant_model_provider inserts in mysql_migration by @jony376 in #15313
• fix(rag): id2image parsing for hyphenated storage object keys (#15117) by @kiannidev in #15118
• fix(api): document image_id parsing for hyphenated thumbnail keys (#15115) by @kiannidev in #15116
• Add Bulk action for linking Multiple Files to Datasets by @Letssharearow in #14960
• Fix: migrate legacy model id configs by @buua436 in #15495
• fix(api): stop duplicating answer in openai-compatible chat completions stream (#15286) by @nickmopen in #15443
• Fix: disable thinking to avoid potential infinite loops in Qwen3.5/Qwen3.6 models by @Aeovy in #15101
• fix: rewrite enable thinking mode for minimax by @Dimon0000000 in #15496
• fix(go-models): remove duplicate roundTripperFunc from novita_test.go by @dripsmvcp in #15492
• feat[Go] implement /api/v1/datasets/<dataset_id>/metadata/config by @Haruko386 in #15493
• Fix: model-provider bugs by @Lynn-Inf in #15460
• fix: post /api/v1/system/tokens by @Hz-186 in #15410
• fix(go-models): harden N1N default transport handling by @glorydavid03023 in #15351
• Feat: mark mysql migrations as applied by @buua436 in #15504
• fix(go-models): use per call context timeouts so long streaming responses are not truncated by @Dexterity104 in #15380
• feat: add Go MCP server update API by @Helios531 in #15261
• Feature: Add MiniMax M3 by @wangq8 in #15513
• Feature: Add the new anthropic and voyage models by @wangq8 in #15516
• Fix: model provider by @Lynn-Inf in #15517
• feat[Go] implement check connection by using apikey and region by @Haruko386 in #15475
• feat[Go]: implement /api/v1/connectors/<connector_id> PATCH by @Dimon0000000 in #15512
• GOCli check provider region by @Hz-186 in #15474
• Fix: model provider orders by @wangq8 in #15524
• [Fix] naive: force-merge short markdown headers to prevent separate chunks by @VictorECDSA in #15488
• Guard missing storage blobs on preview and image endpoints by @kiannidev in #15366
• Fix: Chat completion generation parameter overrides by @bitloi in #15389
• Refa: refine mysql migration version workflow by @buua436 in #15549
• Fix: Model provider add verify and fixed form in modal not resetting issue by @dcc123456 in #15520
• Go: fix gitee balance api by @JinHai-CN in #15554
• Go: refactor provider code by @JinHai-CN in #15564
• Feat: display intl url for siliconflow and verify model provider without llms in json by @Lynn-Inf in #15550
• Fix: agent template smart_customer_service_specialist.json by @buua436 in #15565
• feat[Go]: implement add multi_models by @Haruko386 in #15563
• Chore: update cohere models by @6ba3i in #15576
• Fix: Switching pagesize on a chunk page did not reset the current page. by @dcc123456 in #15401
• Fix: no more @token_required by @wangq8 in #15562
• Go: refactor model provider by @JinHai-CN in #15568
• refactor: overhaul task executor with layered architecture and comprehensive test suite by @xugangqiang in #15471
• Remove unused codes by @JinHai-CN in #15579
• fix(api): guard document rename when linked file row is missing by @bohdansolovie in #15536
• refactor(go-models): add unsupported model driver defaults by @bitloi in #15431
• Fix: consolidate beta auth by @wangq8 in #15581
• feat[Go] implement /connectors/google/oauth by @Haruko386 in #15584
• feat: migrate DELETE /api/v1/datasets/:dataset_id/documents to Go by @xugangqiang in #15577
• fix: show default embedding model when provider is not yet registered by @euvre in #15511
• fix(api): restore DocumentService.accessible check on /preview by @dripsmvcp in #15508
• fix(api): restore accessible check on document preview by @kiannidev in #15505
• Fix: think stream tail handling by @buua436 in #15582
• Fix: search bot and verify model instance by @Lynn-Inf in #15588
• Fix: remove duplicate document preview access check by @buua436 in #15625
• Fix: markdown fenced code block extraction by @bitloi in #15630
• fix: fall back from empty Docling native chunks by @he-yufeng in #15601
• feat: migrate POST /api/v1/datasets/<dataset_id>/documents/stop to Go by @xugangqiang in #15597
• feat: implement GET /api/v1/agents/<agent_id>/versions API by @xugangqiang in #15629
• Fix: model provider by @Lynn-Inf in #15634
• feat: implement POST /api/v1/agents/<agent_id>/upload API by @xugangqiang in #15633
• fix: pass search id to searchbots ask by @buua436 in #15646
• Refactor[Go Model Provider]: refactor baseURL and modelConfig by @Haruko386 in #15627
• fix: duplicate document ingest guard by @buua436 in #15638
• feat: implement GET /api/v1/agents/<agent_id>/versions/<version_id> API by @xugangqiang in #15640
• feat: implement POST /api/v1/searchbots/related_questions API by @xugangqiang in #15639
• feat: migrate meta_filter and convert_conditions to Go by @xugangqiang in #15648
• fix: add nil guard to DocumentDAO.GetByIDs and add tests by @xugangqiang in #15649
• feat(connectors): add Azure Blob Storage data source connector by @JPette1783 in #15466
• refactor: change GetFlattedMetaByKBs return type to common.MetaData by @xugangqiang in #15656
• feat: add ResolveReferenceMetadata utility function by @xugangqiang in #15663
• fix(docker): mount tmpfs on es01 /tmp for entrypoint permissions by @whhe in #15655
• Fix: volc engine and other json key factories by @Lynn-Inf in #15653
• fix(api): set SDK document download Content-Type from filename (#15112) by @kiannidev in #15113
• feat: add KG scoring utilities by @xugangqiang in #15666
• feat: add QueryRewrite prompt builder and response parser by @xugangqiang in #15669
• feat[Go]: implement GET /agents/templates (issue #15240) by @dripsmvcp in #15573
• Fix: preserve markdown tables during delimiter extraction by @bitloi in #15632
• Refactor: consolidate to use @login_required by @wangq8 in #15652
• feat: support SelectFields override in DocEngine for KG-specific queries by @xugangqiang in #15684
• feat: add EnrichChunksWithDocMetadata function to attach document metadata to chunks by @xugangqiang in #15659
• fix: restore case-insensitive contains/not contains/not in and consolidate metadata filter pipeline by @xugangqiang in #15686
• fix: prevent duplicated post-think text by @buua436 in #15651
• feat[Go]: implement Agent/Workflow PUT /api/v1/agents/<canvas_id>/tags by @Haruko386 in #15641
• feat: add KG entity/relation/community search functions by @xugangqiang in #15689
• feat(connectors): add Salesforce CRM data source connector by @JPette1783 in #15462
• feat(go-api): implement MCP server management endpoints by @JPette1783 in #15281
• Go: implement Bedrock embeddings by @tmimmanuel in #15543
• fix: Update Dockerfile and release workflow to use GitHub mirror instead of Gitee by @asiroliu in #15700
• fix(agent): skip empty switch conditions by @he-yufeng in #15691
• Revert "fix(api): authorize owner_ids for list chats and search apps (#14775) by @wangq8 in #15698
• Revert "fix: duplicate document ingest guard" by @wangq8 in #15707
• feat: add KGSearchRetrieval for full KG pipeline (N-hop, scoring, query_rewrite, community) by @xugangqiang in #15690
• Fix: When adding a chat in the main interface, a warning will automatically pop up by @cike8899 in #15685
• feat(go-cli): support batch model add/remove and optional embedding dimension by @Hz-186 in #15631
• feat: Dify-compatible retrieval API endpoint by @xugangqiang in #15704
• fix(onedrive): normalize folder_path for Graph delta URL by @kiannidev in #15503
• Feat: get model list from remote by @Lynn-Inf in #15711
• fix(go-api): sync document handler interface and enforce preview acce… by @Hz-186 in #15688
• Refine handling of POST /api/v1/datasets/search in GO by @qinling0210 in #15583
• fix(retrieval): keep rerank window aligned to page_size for deep pagination by @cleanjunc in #15434
• feat[Go]: implement GET /plugin/tools (issue #15240) by @dripsmvcp in #15570
• fix(rerank): normalize reranker scores onto a single scale before hybrid blend by @cleanjunc in #15429
• Go: implement Xiaomi chat provider by @tmimmanuel in #15626
• fix(go-models): increase JieKouAI SSE scanner buffer by @oktofeesh1 in #15737
• fix: normalize reasoning model families by @bitloi in #15612
• Fix: The time zone is not displayed on the personal profile page. by @cike8899 in #15759
• feat: add placeholder model metas by @buua436 in #15753
• Update Rerank logic in GO by @qinling0210 in #15755
• fix: normalize agent session chunks by @buua436 in #15756
• Fix: The embedded website floating component on the agent page does not display citations. by @cike8899 in #15767
• feat: implement POST /api/v1/searchbots/retrieval_test by @xugangqiang in #15710
• fix(opensearch): keep the BM25 leg in hybrid search by @mateidanut in #15760
• feat[Go]: implement /api/v1/agents/<agent_id>/sessions by @Haruko386 in #15705
• Fix : Ci fail for infinity on level p3 by @6ba3i in #15757
• Addd OpenRouter OpenAI API compatible list models by @wangq8 in #15764
• feat: Unify the 'Add Model Provider' modal by @dcc123456 in #15768
• fix: refine think stream parsing by @buua436 in #15745
• feat: implement FetchChunkVectors for citation vector hydration by @xugangqiang in #15749
• Fix: Model provider bugs by @dcc123456 in #15770
• fix(nlp): tokenize content_tks by whitespace in FulltextQueryer.paragraph by @seekmistar01 in #15721
• Fix: Excel parser emits a spurious header-only chunk at exact chunk_rows multiples by @devYRPauli in #15490
• fix(opensearch): keep "id" in _source on insert so document metadata isn't empty by @ef-rintaro in #15473
• feat(web): show provider count on each model-type filter tag by @sunlight7777777 in #15444
• i18n(it): complete Italian translation (49% → 100%) by @gaulin-ai in #15729
• fix(browser): show resolved variables in workflow run log input by @huang-aoqin in #15325
• feat: add ModelMeta implementations for Xinference, LocalAI, BaiduYiyan, and Tencent Cloud by @euvre in #15752
• fix: support auto mode in table parser document metadata aggregation by @euvre in #15780
• fix: Resolve error when checking pipeline parsing result by @dcc123456 in #15778
• Fix: Add a waiting status to the messages on the chat page. by @cike8899 in #15773
• fix: await lmstudio embedding verification by @buua436 in #15772
• Fix OpenRouter key JSONDecodeError by @wangq8 in #15776
• fix(go-models): fix Moonshot model and balance requests by @oktofeesh1 in #15528
• Go: implement ASR and TTS for xiaomi by @Haruko386 in #15765
• fix(go-models): validate URL suffix config keys by @oktofeesh1 in #15734
• fix(go-models): fix MiniMax driver requests by @oktofeesh1 in #15527
• fix(go-models): normalize Qwen reasoning families by @oktofeesh1 in #15735
• Delete not supported model providers list by @wangq8 in #15783
• feat(ci): enable go test in CI pipeline by @xugangqiang in #15750
• Feat: model list by @Lynn-Inf in #15774
• Fix: is_english() returns False for any list argument (broken language detection) by @devYRPauli in #15489
• fix: guard http.DefaultTransport type assertion in xiaomi for Go 1.25 by @xugangqiang in #15787
• Go: new CLI command, list all models and show model by @JinHai-CN in #15786
• fix(api): guard missing row in SearchService.get_detail by @dearsishs in #15622
• fix(api): await asyncio.wait_for in verify_api_key embedding branch by @dearsishs in #15620
• Fix: Add bge as embedding by @wangq8 in #15784
• Fix: OceanBase tenant startup drift and docs by @6ba3i in #15829
• Fix empty response set not take effect by @wangq8 in #15824
• fix: resolve issue where some models do not use modelInfo parameter by @dcc123456 in #15830
• Fix: list intl Tongyi-Qianwen base_url by @Lynn-Inf in #15831
• Go: refactor CLI by @JinHai-CN in #15728
• feat: Implement checkpoint/resume support for GraphRAG community extraction and entity resolution by @jonathanchang31 in #15523
• Fix: An error message appears when accessing the agent's launch page: "pagesize exceeds maximum value". by @cike8899 in #15835
• Clean CLI for filesystem by @yingfeng in #15838
• feat(api): implement GET /api/v1/agents/prompts endpoint in Go by @Hz-186 in #15748
• fix: propagate memory tenant id in task collect by @buua436 in #15837
• fix: Remove trailing comma in CREATE TABLE tenant_model SQL (#15832) by @Dr1985 in #15836
• Go CLI: update list supported models by @JinHai-CN in #15845
• fix: force image parser json output by @buua436 in #15847
• fix: BaiduYiyan API key validation fails in set_api_key by @euvre in #15828
• Fix: model list by @Lynn-Inf in #15839
• Fix: MinerU cannot be added by @wangq8 in #15841
• fix: prevent None values in auto-metadata from causing KeyError by @xugangqiang in #15842
• Fix VLLM cannot add without /v1 by @wangq8 in #15851
• feat: support custom editing for model list by @dcc123456 in #15855
• fix(models): guard unsafe index access in Google and Ollama drivers by @JPette1783 in #15819
• feat(go-api): Add GET dataset metadata summary API by @Hz-186 in #15843
• fix(go): guard four nil-pointer dereferences causing runtime panics by @JPette1783 in #15815
• Fix: The variables in the Visual Input File of the agent operator are not displayed. by @cike8899 in #15856
• fix(elasticsearch): complete Go result functions by @JSONbored in #15148
• Fix: restore GraphRAG entity ranking by indexing pagerank and n-hop paths by @cleanjunc in #15797
• fix(ci): re-enable Go tests and fix compilation errors after ListModels signature change by @xugangqiang in #15862
• Fix wrong unit test path by @yingfeng in #15864
• feat: implement POST /api/v1/searchbots/ask — streaming RAG with citations and think-tag processing by @xugangqiang in #15825
• feat[Go]: implement /api/v1/agents/<agent_id> and test_db_connection by @Haruko386 in #15771
• feat[Go]: api datasets/<dataset_id>/documents/<document_id>/metadata/… by @Haruko386 in #15846
• Docs: Supported model providers and URLs updated by @writinwaters in #15866
• Go CLI: switch to admin/api server by @JinHai-CN in #15861
• feat(go-api): add RAG retrieval to chat completions by @oktofeesh1 in #15739
• Enhancement: slim Docker image, add .dockerignore, fix Go binary shipping by @6ba3i in #15880
• Fix paddle ocr cannot add by @wangq8 in #15858
• fix: handle qwen rerank error response by @buua436 in #15881
• feat(go): Add DeepSeek models and Gitee alias metadata tests by @Hz-186 in #15885
• Fix: Fix some model provider-related UI issues by @dcc123456 in #15884
• Fix: tenant_model migrate by @Lynn-Inf in #15886
• fix: rename ark_api_key to api_key for volcengine provider config by @dcc123456 in #15896
• Fix add OpenRouter base_url, UI need to select at least one model to verify by @wangq8 in #15894
• Fix: model list by @Lynn-Inf in #15860
• feat: support multi-type factory models by @buua436 in #15893
• Fix: cohere call failing by @6ba3i in #15899
• Go: refactor CLI by @JinHai-CN in #15898
• feat(go-api): migrate agent file download handler to Go with strict P… by @Hz-186 in #15769
• feat[Go]: implement POST /api/v1/files/link-to-datasets by @hunnyboy1217 in #15674
• Fix no need to put inactive models to bottom by @wangq8 in #15903
• fix: remove think text buffering for real-time reasoning stream by @shaoqing404 in #15891
• Fix cannot add bedrock by @wangq8 in #15904
• Fix: model list by @Lynn-Inf in #15905
• fix: strip Ollama-style tag suffix from LocalAI model names by @xugangqiang in #15908
• fix: route visual agent calls to image model by @buua436 in #15906
• Fix: The dataset retrieval test returned an incorrect total number. by @cike8899 in #15901
• Fix: add thin scrollbar styling for x-spreadsheet component by @dcc123456 in #15912
• Fix cannot add fish audio by @wangq8 in #15913
• Fix: add image2text/speech2text/ocr support by @dcc123456 in #15915
• fix(go): Fix case-insensitive model alias lookup by @Hz-186 in #15911
• feat(go): Update Qwen models in all_models.json by @Hz-186 in #15910
• Go: fix test and remove unused code by @JinHai-CN in #15909
• feat: Implement API of ragflow server in Go by @jonathanchang31 in #15256
• feat(go-api): port forgot-password flow to Go (#15282) by @Rene0422 in #15290
• Fix: model list by @Lynn-Inf in #15914
• Fix multiple tags by @wangq8 in #15931
• fix(deps): exclude litellm 1.82.6 (internal ImportError) — #15916 by @Rene0422 in #15920
• fix: allow zero message history window and clear history for new sessions (#15897) by @zaviermeekz-cpu in #15902
• fix(sync): tolerate list inputs for Discord server_ids / channels (#15790) by @Rene0422 in #15809
• fix(go-models): guard nil pointers in DeepSeek and VolcEngine streaming by @JPette1783 in #15817
• Go: Refactor list-models func by @Haruko386 in #15900
• Go: add parse and chunk command by @JinHai-CN in #15936
• Fix Agent chat Minimax content in thinking by @wangq8 in #15937
• Fix: The regular expression configuration for pipeline header-based chunking will be reset. by @cike8899 in #15935
• fix: add model_name to embed completion request (#15883) by @zaviermeekz-cpu in #15888
• fix(es): downgrade LLM-generated invalid SQL to WARNING in ES sql() (#15409) by @Rene0422 in #15709
• fix: replicate model provider by @6ba3i in #15933
• fix(api): guard dataset delete when File2Document row is missing by @bohdansolovie in #15533
• fix(preview): authenticate markdown document preview requests by @monsterdavidliu-ux in #15589
• fix(db): drop Peewee-auto-named unique index on tenant_model_instance (#15699) by @Rene0422 in #15879
• fix(connectors): enforce WebDAV numeric string size limits by @oktofeesh1 in #15731
• fix(api): POST /documents/stop removes partial chunks and resets counters by @kiannidev in #15789
• fix(dialog): guard async_ask() against empty or invalid kb_ids by @bohdansolovie in #15530
• fix: tolerate raw api_key string in AzureEmbed and AzureGptV4 init by @jaso0n0818 in #15877
• fix: guard docStoreConn.delete with index_exist in parse and stop_parsing by @jaso0n0818 in #15876
• Fix: Remove the pagination from the search and retrieval pages. by @cike8899 in #15942
• i18n(ko): add Korean (한국어) translation by @yooxdb in #15863
• fix: propagate max_tokens from model config to downstream consumers by @shaoqing404 in #15945
• Fix: get is_tools from model record by @Lynn-Inf in #15946
• feat(go): Add embedding dimension metadata and validation by @Hz-186 in #15939
• Fix: Search mindmap not working by @wangq8 in #15949
• Docs: Added initial draft for v0.26.0 release notes. by @writinwaters in #15603
• Docs: Update version references to v0.26.0 in READMEs and docs by @asiroliu in #15941

New Contributors

• @Helios531 made their first contribution in #15131
• @leslie2046 made their first contribution in #15383
• @monsterdavidliu-ux made their first contribution in #15386
• @Dexterity104 made their first contribution in #15382
• @galuis116 made their first contribution in #15181
• @ef-rintaro made their first contribution in #15390
• @davidmyriel made their first contribution in #15361
• @glorysr1209-png made their first contribution in #15359
• @Aeovy made their first contribution in #15101
• @VictorECDSA made their first contribution in #15488
• @he-yufeng made their first contribution in #15601
• @cleanjunc made their first contribution in #15434
• @mateidanut made their first contribution in #15760
• @seekmistar01 made their first contribution in #15721
• @devYRPauli made their first contribution in #15490
• @sunlight7777777 made their first contribution in #15444
• @gaulin-ai made their first contribution in #15729
• @dearsishs made their first contribution in #15622
• @Dr1985 made their first contribution in #15836
• @zaviermeekz-cpu made their first contribution in #15902
• @jaso0n0818 made their first contribution in #15877
• @yooxdb made their first contribution in #15863

Full Changelog: v0.25.6...v0.26.0