nightly

Release nightly created from f45f03a at 2026-07-20 21:40:32+08:00

v0.26.4

Summary

Released on July 7, 2026.

New features

• NLP/Tokenization: Adds a language-aware Snowball stemmer supporting 16 languages, integrates the dataset language parameter across the tokenization pipeline, and adds Dutch to the frontend. #14140

Bug fixes

• The system crashed with a ValueError when parsing LM-Studio model names containing an '@' symbol. #16467
• The MCP server crashed because the list_chats function expected a list from the /chats API but received a paginated dictionary instead. #16639
• The Docling parser silently dropped mathematical formulas from documents instead of extracting them. #16645
• The system failed to persist inline edits made to metadata values to the backend. #16655
• The system removed existing links when bulk-linking files to datasets. #16587
• The filter failed to use Chinese. #16673
• The system failed to enforce that tag weights must be greater than zero. #16657
• The layout of the agent prompt dropdown menu rendered incorrectly. #16653
• The Laws chunker incorrectly classified dotted-number cross-references as headings. #16626
• The system failed to prevent users from saving duplicate MCP server names during the editing process. #16588
• The system allowed users to export an empty MCP server selection. #16589

What's Changed

• Go: unify three services into one binary by @JinHai-CN in #16462
• Feat: Add knowledge compilation workflows by @KevinHuSh in #16515
• Go: file syncer service framework by @JinHai-CN in #16579
• fix(agent/tools): port AkShare to ToolBase so it works as an Agent tool by @immuhammadfurqan in #16417
• fix: prevent duplicate 'skills' and '.knowledgebase' folders caused by race conditions by @euvre in #16568
• Fix harness DAG slow-branch test cased by nil initialization of pregel engine by @yingfeng in #16591
• Refactor: reformat all code for lefthook using ruff and gofmt by @wangq8 in #16585
• fix: preserve existing links when bulk linking files to knowledge bases by @euvre in #16587
• fix: agent chat completions can not use by @Haruko386 in #16570
• fix: unable to import mcp from local by @Haruko386 in #16590
• fix: unable to fetch tools for MCP by @Haruko386 in #16583
• fix(go-file): KB counter drift when deleting files with linked documents by @Hz-186 in #16584
• fix: prevent exporting empty MCP server selection by @euvre in #16589
• fix: prevent duplicate MCP server name when editing by @euvre in #16588
• Go: fix 'list services' by @JinHai-CN in #16598
• i18n: localize visual input file label in agent form by @euvre in #16594
• Fix: optimize folder data handling in MoveDialog component by @dcc123456 in #16580
• Go: refactor by @JinHai-CN in #16602
• fix(agent/tools): port Crawler to ToolBase so it can load and run by @immuhammadfurqan in #16415
• Refactor(harness): remove naive inline graph engine , unify graph execution under single pregel engine by @yingfeng in #16608
• Port agent PRs to GO - 3 by @qinling0210 in #16596
• Feat: add gbrain compile template for session/memory data by @yingfeng in #16613
• fix: get all memory in team with permission=me by @Haruko386 in #16593
• fix: get team merber's chat by @Haruko386 in #16597
• fix: get team's search in own search-list by @Haruko386 in #16599
• Fix Go: list providers order issue. by @JinHai-CN in #16616
• fix: cannot get query in agent-log by @Haruko386 in #16610
• fix(agent): defend against @ in var names at all template-split sites by @skbs-eng in #16469
• Go: Update error info by @JinHai-CN in #16619
• Go:cli move _order _columns sort group by @myf-bee in #16615
• fix(agent): enable MCP file preview via doc_id by @monsterdavidliu-ux in #15399
• docs: update Docker build instructions for deps image by @asiroliu in #16620
• Update development script by @wangq8 in #16623
• Feat: Put some wiki templates. by @KevinHuSh in #16617
• Delete canvas_app.py and evaluation_service.py by @wangq8 in #16614
• Cap RAPTOR UMAP n_neighbors to prevent OOM on large datasets by @afrozenpeach in #16627
• Align Go ingestion boundaries with Python by @yuzhichang in #16647
• Refactor: pdf parser by @xugangqiang in #16625
• fix(agent): collect CodeExec artifacts from ReAct tool responses by @euvre in #16609
• fix(agent): enable single-component debug for Agent in Go backend by @euvre in #16606
• fix(go-skill): Elasticsearch skill search field mapping by @Hz-186 in #16611
• fix(go-document): add document and file access checks by @Hz-186 in #16592
• feat(agent): expose thinking mode control per LLM node in Agent canvas by @AI-Mart in #16640
• Switch the default minio image in the helm chart as well by @carlobeltrame in #16322
• Port agent PRs to GO - 4 by @qinling0210 in #16652
• fix: Docling parser extracts mathematical formulas by @Sohaib9920 in #16645
• fix(api): cap memory message limit and top_n at REST_API_MAX_PAGE_SIZE by @jony376 in #15376
• Dev: Fix nats host by @wangq8 in #16656
• fix(nlp): prevent dotted-number cross-references from being classified as headings in Laws chunker by @OSHA-B in #16626
• fix(metadata): inline value edits not persisted to backend by @dcc123456 in #16655
• fix(go-agent): align agent debug input form with Python by @Hz-186 in #16654
• feat(go): add length split strategy with overlap to chunk pipeline by @Crystora in #16047
• feat(go): drop empty and duplicate chunks in postprocess filter by @Crystora in #16049
• Fix OpenAI response created timestamp by @Harsh23Kashyap in #16401
• Fix: Layout of the agent prompt dropdown menu is messed up. by @cike8899 in #16653
• Fix Tag weight should be greater than 0 by @wangq8 in #16657
• Fix Agent session lost tag by @wangq8 in #16670
• Go: merge functions by @JinHai-CN in #16622
• Fix filter to use Chinese by @wangq8 in #16673
• fix(go): uploaded documents should be enabled by default by @euvre in #16674
• Refactor: Remove redundant functions - and fix run_raptor to do_raptor by @KevinHuSh in #16671
• fix(go-agent): Yahoofinance input and run by @Hz-186 in #16658
• fix(go-agent): missing input form for ExeSQL and Browser agent nodes by @Hz-186 in #16675
• fix(go-document): keep upload partial success data as array by @Hz-186 in #16661
• fix(go): add missing 'resume' chunk method for new tenants by @euvre in #16660
• Align Go parser backends and PDF pipeline with Python by @yuzhichang in #16676
• fix(graphrag): filter negative-judgment and misattributed relationship edges by @afrozenpeach in #16541
• fix: handle missing ES/OpenSearch ind...

v0.26.3

Summary

Released on July 2, 2026.

New features

• Data source connector: Introduces Google BigQuery as a data source connector for document ingestion and incremental syncing. #15871
• MCP: Adds two MCP tools ragflow_list_datasets and ragflow_list_chats to the RAGFlow MCP server. #15384
• File ingestion
  • Integrates the layout-aware SoMark OCR parser into the system for extracting and indexing complex document elements like tables and figures. #16482
  • API: Exposes the Ingest documents endpoint for processing documents using a customized pipeline. #16395

Improvements

• Backend: Introduces partial success handling for batch document uploads, ensuring that a single failed file no longer causes the entire batch to be discarded. #16438
• UI/UX: Refactors the global header to dynamically switch between a full desktop navigation bar and a mobile-friendly side drawer based on available screen space, resolving overlapping issues and significantly improving the mobile user experience. #15984

Bug fixes

• After setting up the Ollama provider, the system returned the error: 'Failed to access model(name) using this api key'. #16519
• Deleting a user/assistant chat message pair mistakenly removed the previous conversation turn's reference due to an index misalignment. #16436

What's Changed

• refactor: enhance UI components and improve layout by @carlh7777 in #15984
• fix: update variable completeness check to allow None parameter by @buua436 in #16389
• Go CLI: Add create and drop commands by @JinHai-CN in #16430
• Docs: Updated release date and cli installation commands by @writinwaters in #16435
• Fix PR template by @JinHai-CN in #16439
• feat: handle partial upload success in document batch upload by @euvre in #16438
• Guard /datasets/{dataset_id}/chunks cannot parse ingestion pipeline, use /documents/ingest instead by @wangq8 in #16395
• fix(rag/nlp): handle non-numbered DOCX heading styles by @Harsh23Kashyap in #16219
• Fix: failed to get embedding model by embd_id: model config not found BAAI/bge-m3@...@siliconflow by @wangq8 in #16445
• fix: user-setting modal fixes and DOMPurify cleanup by @dcc123456 in #16449
• [Go] Fix searchbot BETA auth by @wangq8 in #16450
• Go CLI: refactor commands by @JinHai-CN in #16447
• [Go] Fix searchbot retrieval_test accept kb_id as array, fix model recognize by @wangq8 in #16452
• Refactor: oss parser go refactor by @xugangqiang in #16391
• fix: auth middleware double responses on early rejection by @Haruko386 in #16444
• fix: unable to upload avatar for search by @Haruko386 in #16437
• fix: new chat cannot be edit by @Haruko386 in #16434
• fix(py): chat message reference deletion index by @Hz-186 in #16436
• feat(go-api): Add Go chat session message delete and feedback APIs by @Hz-186 in #16442
• [Go] Fix beta auth for /documents/images/:image_id and /documents/:id/preview and /thumbnails by @wangq8 in #16453
• Go CLI: fix api commands by @JinHai-CN in #16457
• feat[Go]: implement searches/<search_id>/completions POST by @Haruko386 in #16440
• Docs: Added an FAQ by @writinwaters in #16466
• Revert "feat(go-api): Add Go chat session message delete and feedback APIs" by @JinHai-CN in #16465
• Feature big query connector by @Attili-sys in #15871
• Go CLI: add show users plan summary by @JinHai-CN in #16463
• Go CLI: add admin_command response table funtion by @myf-bee in #16454
• build(deps): bump crawl4ai from 0.8.9 to 0.9.0 by @dependabot[bot] in #16470
• Fix: allow any host for url for development by @wangq8 in #16459
• fix: adjust width of messageItemSectionLeft to fit-content by @dcc123456 in #16488
• Go CLI: merge function by @JinHai-CN in #16458
• fix(agent): inject uploaded attachments into LLM context (#15215) by @Rene0422 in #15220
• Port 14 upstream agent security / correctness fixes to Go canvas by @yuzhichang in #16455
• [Go] Add /api/v1/searchbots/mindmap and /api/v1/chat/mindmap by @wangq8 in #16443
• fix(agent): return session_id when chat completion produces no events (#15169) by @Rene0422 in #15228
• fix: remove dup-prefix in bot_routes by @Haruko386 in #16492
• Fix: UI cannot start up by @wangq8 in #16497
• fix(agent): add canvas_type filter and field to list_agents API by @huang-aoqin in #15754
• Fix harness streaming emit by @yingfeng in #16486
• feat(go-api): Migrate Box web OAuth connector APIs to Go by @Hz-186 in #16480
• Add spacy based ner and relationship extractor for both python and Go version with equivalent outputs by @yingfeng in #16456
• revert: roll back tests.yml CI changes from PR #16391 by @xugangqiang in #16505
• Fix: CI by @Lynn-Inf in #16504
• fix(common/time_utils): correct None/empty timestamp fallback and ISO 8601 parsing by @Harsh23Kashyap in #16483
• fix(agent): prevent empty LLM user message after prompt fitting by @Taranum01 in #16413
• fix: handle non-serializable objects in agent canvas SSE and state se… by @RazmikGevorgyan in #14210
• Add CAJAL scientific paper agent template by @sxxtony in #14641
• feat(mcp): add ragflow_list_datasets and ragflow_list_chats tools by @saltsalt123 in #15384
• Fix g++ 11 incompatibility issue by @yingfeng in #16512
• fix(agent/tools): GoogleScholar empty json output and ignored top_n by @immuhammadfurqan in #16419
• [Go] Add API /api/v1/chat/recommendation and consolidate with /api/v1/searchbots/related_questions by @wangq8 in #16500
• Fix: ollama provider by @Lynn-Inf in #16519
• Feat: SoMark by @Lynn-Inf in #16482
• fix(web): sanitize agent rerun modal HTML against stored XSS by @Harsh23Kashyap in #16516
• fix(agent): filter TuShare news with upstream keyword input by @Harsh23Kashyap in #16361
• feat(agent): add BGPT structured literature evidence search tool by @connerlambden in #16050
• fix(deepdoc): parse bodyless HTML fragments by @Harsh23Kashyap in #16423
• fix(qa): preserve final CSV pair row number by @Harsh23Kashyap in #16433
• Implement chat completions in go by @qinling0210 in #16491
• fix(api): fall back to factory max_tokens for tenant models by @Harsh23Kashyap in #16364
• fix(harness): truncate text on rune boundary to keep UTF-8 valid by @feizhuzheng in #16511
• feat(frontend): add AuthenticatedImg component for authorized image requests by @euvre in #16525
• Fix GetProjectRoot in GO by @qinling0210 in #16520
• refactor: replace context.WithCancel with t.Context by @blackflytech in #16509
• fix: unable to open filter in agent page(no agent tags...) by @Haruko386 in #16531
• Go CLI: fix list provider models by @JinHai-CN in #16493
• fix: normalize Q&A parser ID key to lowercase 'qa' by @euvre in #16530
• fix: return call failed when LLM not ava...

v0.26.2

Summary

Released on June 29, 2026.

New features

• Chat channels:
  • Integrates WhatsApp via a QR code scan. #16238
  • Integrates DingTalk via bot API credentials. #16183
  • Integrates WeCom via WebSocket connection. #16175
• File parsing: Adds fallback logic for PP-OCRv6 and similar text recognition models, and integrates image parsing within the PaddleOCR pipeline. #16150

Improvements

• Datasets
  • Adds an edge-case check to the file ingestion logic to safely link empty folders to a dataset. #16296
  • Adds pagination to the get_flatted_meta_by_kbs() method to prevent CRUD failures relating to datasets with more than 10,000 documents. #16095
• Chat channels: Ensures end-user conversation histories persist across restarts but still separate when a channel is bound to a new dialog. #16274

i18n

• Expands i18n coverage and refines translations for the Korean UI. #16203；
• Adds missing French translations for chat channels, username validation, and model editing. #16217

Bug fixes

• File parsing:
  • A ValueError was raised during .docx parsing on 'Heading' styles. #16284
  • Tables in .docx files were silently removed by the laws document parser. #16155
  • Lone Markdown headers were separated from their subsequent sections when a DeepDoc delimiter was set. #16109
  • Markdown tables appeared twice in chunk results, once as a standalone chunk and again within a regular text chunk. #16143
• MCP
  • The MCP server hung indefinitely when encountering empty document pages; the final page of documents was silently dropped due to flawed pagination. #16285
  • The MCP connection and dataset discovery process failed because the server's hardcoded fetch limit exceeded the introduced page_size restriction. #16148
• Dataflow
  • Restores the dataflow rerun endpoint and ensures the ingestion response includes the DSL payload to reliably render the timeline and parser views. #16292
  • Restores dataflow parser defaults and returns the actual SSE payload, ensuring pipeline runs correctly surface message IDs and log updates. #16290
• LLM: Failed to use the new gemini-3.5-flash model via the Google Cloud (Vertex AI) provider using eu or us region endpoint. This is resolved by explicitly multi-region routing requests to the required aiplatform.<region>.rep.googleapis.com domains instead of synthesized <region>-aiplatform.googleapis.com hosts. #15990
• UI/UX: The metadata add modal sent empty values to the backend. #15229

What's Changed

• Fix the launch script by @wangq8 in #16159
• Refactor(go-cli): improve variable and label naming in CLI parseAddModel by @Hz-186 in #16145
• Fix Chat/Search/Agent bot show image by @wangq8 in #16152
• Fix: MCP dataset discovery fails due to REST API max page size limit … by @xuhaiL in #16148
• Fix: cli points to the wrong api endpoints by @6ba3i in #16171
• feat(go-api): implement Go-side document PATCH API & align parsing/metadata sync behavior by @Hz-186 in #15975
• feat(go-api): migrate MCP server detail and download API to Go by @Hz-186 in #16113
• feat: support wecom websocket channel by @buua436 in #16175
• Fix: Lint error. by @cike8899 in #16172
• fix: fix some bugs in check_conn and drop_inst by @Haruko386 in #16180
• fix(deepdoc): attach lone header lines to the following section when delimiter is set by @jaso0n0818 in #16109
• Go CLI: add support for windows, linux, macos by @Haruko386 in #16082
• Go: add statistics command by @JinHai-CN in #16119
• Fix search rename error with multiple error message (#664) by @wangq8 in #16186
• Go CLI: add support for windows, linux, macos by @Haruko386 in #16184
• fix: use RESTful pipeline detail API with knowledgeId and logId by @euvre in #16182
• Fix: replace tenant_llm apis by @Lynn-Inf in #16131
• fix: add nightly tags by @Haruko386 in #16189
• Go: add command, show tasks summary by @JinHai-CN in #16187
• fix: chat channel defaults and feishu shutdown by @buua436 in #16176
• fix: wrong clang/toolchain for windows by @Haruko386 in #16191
• Go: add command, list, remove, stop tasks by @JinHai-CN in #16190
• feat[Go]: add modelID for delete_model and update_status by @Haruko386 in #16025
• feat[go]: implement /api/v1/datasets/<dataset_id> PUT by @Haruko386 in #16122
• feat[Go]: implement datasets/<dataset_id>/index P/G by @Haruko386 in #16153
• Json add paddleOCR models by @Haruko386 in #16156
• Implement OpenAI chat completions in GO by @qinling0210 in #16177
• fix: hard-coded paths for Windows C compiler by @Haruko386 in #16193
• Fix that agent cannot be the same name by @wangq8 in #16192
• feat: add dingtalk chat channel by @buua436 in #16183
• Go: fix duplicate models by @JinHai-CN in #16197
• Go CLI: add list user commands framework by @JinHai-CN in #16201
• Go CLI: refactor some commands by @JinHai-CN in #16204
• Go CLI: refactor commands by @JinHai-CN in #16213
• Go CLI: admin list configs by @JinHai-CN in #16221
• Fix: preserve tables when parsing DOCX with the laws parser (#16008) by @manan-tech in #16155
• fix: Enhance Windows build process for office_oxide and rag tokenizer by @Haruko386 in #16223
• feat(agent): align Go agent behavior with Python (except retrieval component) by @yuzhichang in #16225
• fix: Enhance Windows build configuration in release.yml by @Haruko386 in #16227
• Fix: metadata add modal sends empty value due to stale closure by @nickmopen in #15229
• Go CLI: Fix alter role by @JinHai-CN in #16226
• fix: fix release workflow for Windows and MAC builds by @Haruko386 in #16235
• feat[Go]: implement api /api/v1/datasets/<dataset_id>/chunks POST by @Haruko386 in #16067
• fix: unable to chat after set model by @Haruko386 in #16195
• feat(go-api): complete chat channel API migration with tests by @Hz-186 in #16139
• feat(go-api): Implement BulkDeleteChats Go API and fix ListChats by @Hz-186 in #16157
• feat(go-api): migrate searchbot share detail endpoint to go by @Hz-186 in #16124
• Add show / set role default models by @JinHai-CN in #16240
• Feat: Add a prefix to the name of the FormField associated with the chat. by @cike8899 in #16178
• i18n(fr): add missing French translations for chat channels, username validation and model editing by @vincentlambert in #16217
• test(go): ensure go unit tests pass by @yuzhichang in #16241
• Go CLI: admin list providers by @JinHai-CN in #16243
• fix: handle SIMDe headers installation for arm64 by @Haruko386 in #16244
• Fix: When re-entering the agent page, the data from the previous session flashes briefly. by @cike8899 in #16251
• fix: propagate contextvars through thread_pool_exec by @vincentlambert in https://github.com/infiniflow/rag...

v0.26.1

Summary

New features

• Model provider: Allows users to modify the model type of existing model configurations. #16029
• Chat channel: Enables users to deploy RAGFlow assistants as chatbots within external messaging platforms like Discord and Feishu. #15850
• Observability: Groups multi-turn chat traces by session within Langfuse to facilitate conversation debugging and replay. #15679

i18n

• Adds ~70 missing translation keys for the French UI. #15983

Bug fixes

• Model provider: Fixes inaccurate token billing, corrects embedding truncation limits to prevent indexing failures, and refactors error handling to surface root error causes. #15424
• API: Adds a legacy mode to the /chat/completions endpoint to restore the v0.23.0 style <think> output for backward compatibility. #16014 See also Converse with chat assistant.

What's Changed

• Go: add dimensions for list models and fix some embed-bug in providers by @Haruko386 in #15940
• feat(models): add shared HTTP client, SSE parser, and stub helpers for Go model drivers by @JPette1783 in #15821
• fix(user-settings): collapse sidebar to icon-only rail on mobile by @carlh7777 in #15678
• fix(llm): correct error handling, token accounting, and truncation in embedding providers by @Dexterity104 in #15424
• Improve docx preview by @yingfeng in #15907
• fix: keep max pagerank for repeated n-hop edges by @he-yufeng in #15696
• feat: Langfuse session grouping for multi-turn chat traces by @jonathanchang31 in #15679
• fix(profile): enforce profile name validation and input constraints by @carlh7777 in #15694
• Json: add some models by @Haruko386 in #15947
• feat(go): Add Z.ai model entries to all_models.json Add missing Qwen commercial models and provider aliases by @Hz-186 in #15929
• Go: use NATS as the message queue by @JinHai-CN in #15327
• fix(go): redact internal handler errors by @bitloi in #15746
• Fix: chat/agent -- Default avatar is not displaying correctly. by @cike8899 in #15948
• Go: parse ingestion DSL by @JinHai-CN in #15938
• Decouple skill space from Python API by @yingfeng in #15971
• Feat: chat channels — connect assistants to external messaging bots by @KevinHuSh in #15850
• fix(Go-models): preserve model name lookup when aliases exist by @Haruko386 in #15969
• Json[model-provider] add nvidia, moonshot, minimax, claude, GPT models by @Haruko386 in #15970
• Go: add office_oxide and parse docx file. by @JinHai-CN in #15976
• fix: SSE write timeout by @bitloi in #15852
• feat(agent): Go port — canvas engine, 22 components, DSL v2, 13 endpoints by @yuzhichang in #15952
• Go: add more file parser by @JinHai-CN in #15979
• i18n(fr): add ~70 missing French translation keys by @vincentlambert in #15983
• Go: file parser config by @JinHai-CN in #15989
• Fix last login time by @wangq8 in #16004
• Go: refactor model API to accept model id by @JinHai-CN in #15999
• fix: use /api/tags endpoint for Ollama model listing (#16000) by @zaviermeekz-cpu in #16003
• Add git-like file commit API by @yingfeng in #15978
• feat(go-models): Add google models to all_models.json by @Hz-186 in #16007
• feat: add custom value support for s3 region by @buua436 in #15968
• refactor(go): replace GenerateUUID1 with GenerateToken for entity IDs by @Hz-186 in #16010
• Move REDIS to engine dir by @JinHai-CN in #16006
• Go: add Markdown parser by @JinHai-CN in #16016
• Json: add many models to all_models.json by @Haruko386 in #16013
• fix: output 2 lines when list-supported models by @Haruko386 in #16015
• fix(api): return 4xx not 500 when attachment blob is missing by @dripsmvcp in #15509
• Go: parse HTML file by @JinHai-CN in #16018
• Go: add API mode check in file system command by @JinHai-CN in #16022
• Fix: Adjust chat line height. by @cike8899 in #16021
• Fix one data source can be synced to multiple dataset by @wangq8 in #16023
• Feat: Move less important chat settings into a collapsible panel. by @cike8899 in #16024
• Feat: model provider by @Lynn-Inf in #16028
• Feat: Add edit model type function by @dcc123456 in #16029
• Feat: add harness-go framework —— graph engine by @yingfeng in #16039
• feat[Go] add max_dimension and dimensions for ModelRequest by @Haruko386 in #16019
• fix: add legacy chat/completions mode by @buua436 in #16014
• fix: sandbox config api method mismatch by @buua436 in #16031
• Feat: add harness-go framework —— agent core by @yingfeng in #16045
• Dev: consoldiate dev script by @wangq8 in #16066
• Refactor: Change table chat_channel status data type. by @KevinHuSh in #16061
• feat(go-api): support setting tenant default models by model_id by @Hz-186 in #16030
• feat(go-cli): support add models with embedding type by @Hz-186 in #16020
• feat(go-models): add more providers by @Hz-186 in #16017
• fix: improve remember me checkbox UX in login form by @BasilFoubert in #16051
• Fix: filter unsupported model_type by @Lynn-Inf in #16062
• fix: add toc field to extractor output by @buua436 in #16059
• Go: fix register user by @JinHai-CN in #16058
• Document: github release for RAGFlow Go CLI by @Haruko386 in #16036
• fix(go): all_models.json by @Hz-186 in #16075
• Fix: v0.26.1 model provider by @Lynn-Inf in #16073
• fix: resolve the error caused by office_oxide by @Haruko386 in #16078
• Fix auto migration issue by @JinHai-CN in #16081
• fix: update localization keys for image2text and add ocr option by @dcc123456 in #16076
• fix: update channelTemplates to filter for Discord and Lark only by @dcc123456 in #16065
• fix: remove unnecessary div in profile page layout by @dcc123456 in #16091
• fix workflow ss not found by @myf-bee in #16085
• fix: remove unnecessary 'asChild' prop from FilterButton component by @dcc123456 in #16094
• Docs: Added v0.26.1 release notes by @writinwaters in #16087
• Fix: SSRF in markdown parser remote image fetch by @galuis116 in #15438
• fix: rename dialog_id to chat_id in chat_channel (backend + frontend) by @KevinHuSh in #16096
• Refactor: rename /chat_channels to /chat-channels by @wangq8 in #16099
• build(deps): bump starlette from 0.51.0 to 1.3.1 by @dependabot[bot] in #16089
• Go refactor: merge similar functions by @JinHai-CN in #16098
• refactor(paddleocr): migrate from sync API to async Job API by @Rander7 in #15967
• feat(paddleocr): add image parsing support with async Job API by @Rander7 in #16086
• fix: paginate non-DeepDOC PDF pars...

v0.26.0

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 https://github.com/infi...

v0.25.6

Summary

Released on May 26, 2026.

New features

• Agent: Adds a Browser component that enables AI to autonomously navigate and interact with web pages. #14888

Improvements

• RAG: RAPTOR construction now introduces AHC mode (Ψ-RAG), which expands semantics from the document level to the dataset level. Not only is index construction performance significantly higher than the previous RAPTOR, but it also outperforms the old RAPTOR on Recall@5 and average F1. Users can switch between AHC and GMM modes. This feature was introduced in v0.25.3, and this release fixes several bugs. #14674#14679
• Agent: Introduces lightweight @tool decorator to streamline Python function registration process for chat models. #15047
• Agent: Enables agent messages to display base64-encoded images. #15212
• Agent: Exposes Doc Generator component's file metadata as discrete variables. #15080
• Agent: Allows developers to pass chat_template_kwargs to agent chat completion endpoint. #14182#14542

Bug fixes

• Fixes /chat/completions to allow sending only latest message in API payload and removes requirement to transmit full conversation history. #15197 See also Converse with chat assistant.
• Weight assigned to vector similarity was not applied during the retrieval phase. #15108
• Fixes parser configs not saving on dataset configuration page.
• Logs wer not fully displayed on data source details page.
• Fixes document status filtering failure.
• Fixes crash guard for empty LLM choices responses.
• RAG: RAPTOR construction process halted when using the Infinity document engine. #14998
• Fixes streaming response parsing for Mistral/Upstage reasoning models.
• Fixes HTML tags in ingestion pipeline parser output.
• Fixes table parser metadata.
• Fixes asyncio event loop nesting and fire-and-forget task issues.
• Fixes asyncio.Semaphore bound to different event loop error.
• Agent: Fixes Agent component prompt variable disappearing and search vector_similarity_weight issues.

What's Changed

• Update chunk/metadata cli by @qinling0210 in #15055
• Fix: The logs on the data source details page are not fully displayed. by @dcc123456 in #15056
• Go: implement Embed (embeddings) in TogetherAI driver by @jack-stormentswe in #15017
• Go: implement provider: n1n.ai by @sxxtony in #15010
• Go: implement Rerank in Xinference driver by @RenzoMXD in #15032
• Go: implement Rerank in Novita driver by @RenzoMXD in #15014
• Go: implement Embed in Xinference driver by @RenzoMXD in #14932
• Go: implement provider: GPUStack (chat) by @RenzoMXD in #15024
• feat(go-models): add Azure OpenAI model driver by @web-dev0521 in #15022
• fix: prevent session user_id spoofing via request body by @enjoyandlove in #15077
• fix(api): check kb ownership in /dify/retrieval by @dripsmvcp in #15028
• Go: implement ASR and TTS for Xinference by @Haruko386 in #15096
• Docs: update python version to 3.13 by @JinHai-CN in #15103
• Fix: <asyncio.locks.Semaphore object at 0xabcd [locked]> is bound to a different event loop by @wangq8 in #15100
• Go: implement rerank, asr, tts for TogetherAI by @Haruko386 in #15107
• Docs: Updated v0.25.5 release notes by @writinwaters in #15109
• Docs: Fixed a deployment issue by @writinwaters in #15114
• Revert "fix(api): infer /documents/{id}/download Content-Type from filename when ext is omitted (#15052)" by @buua436 in #15138
• Fix: guard missing task language by @buua436 in #15136
• feat(go-models): add PPIO provider driver by @JSONbored in #15099
• Go: implement ASR in ZhipuAI driver by @jakearmstrong59 in #15134
• fix: block SSRF in misc_utils.download_img for OAuth avatars by @dale053 in #14868
• feat: Add SDK and cURL examples for chunk management, chat assistant, and retrieval (#4310) by @bhongong in #14208
• Refactor: enahnce retry and timeout by @wangq8 in #14983
• Refactor: enahnce CI by @wangq8 in #15147
• fix: prevent sensitive fields from leaking in user API responses by @dale053 in #14792
• feat: pass chat_template_kwargs through agent chat completion by @jsdevninja in #14542
• Go: implement ASR in OpenRouter driver by @jakearmstrong59 in #15067
• feat(evaluation): track token usage in evaluation results by @kiannidev in #13487
• Go: implement provider: TokenPony by @dripsmvcp in #15091
• fix: move agent attachment download api by @buua436 in #15146
• Fix: add model_type into llm_setting by @Lynn-Inf in #15141
• feat(go-models): add Groq provider driver by @JSONbored in #15097
• Fix search vector_similarity_weight by @wangq8 in #15108
• Refactor: Move API files by @wangq8 in #15151
• Docs: Added a guide on integrating Discord. by @writinwaters in #15156
• Go: implement embed, rerank, tts for AstraFlow by @Haruko386 in #15135
• Go: implement reasoning_chat, TTS, ASR for Groq by @Ltohka in #15153
• Fix: /openai/<chat_id>/chat/completions not aware of session_id by @wangq8 in #15155
• fix(go): support OpenAI audio endpoints by @bitloi in #15104
• feat(i18n): complete French translation — add ~1400 missing keys by @vincentlambert in #15192
• Fix: /chat/completions not aware of conversation_id by @wangq8 in #15162
• Go: implement provider: Tencent Hunyuan by @dripsmvcp in #15092
• Go: implement provider: PaddleOCR_Local by @Haruko386 in #15158
• Go: add ingestion server by @JinHai-CN in #15094
• Revert "Fix: /openai/<chat_id>/chat/completions not aware of session_id" by @wangq8 in #15205
• Fix /chat/completions to allow send only the latest message by @wangq8 in #15197
• feat[go]: implement provider: TokenHub by @Dimon0000000 in #15159
• Go: implement embed for Tencent Hunyuan by @Haruko386 in #15207
• Feat: Expose Doc Generator file metadata as discrete outputs by @nickmopen in #15080
• Fix: table parser metadata by @ahmadintisar in #15127
• fix: The output of the parser in the ingestion pipeline contains HTML tags by @jonathanchang31 in #14920
• Fix: Correct the API path by @wangq8 in #15204
• Fix: Replace the red highlight at the top of the PDF document with yellow. by @cike8899 in #15203
• fix(Go): rewrite chat, listmodels, embed for Ollama by @Haruko386 in #15213
• Fix #15170 cannot filter document status by @wangq8 in #15216
• Feat: Enable agent messages to display base64 images by @cike8899 in #15212
• Feat: add new tests and tescases for restful api suite by @6ba3i in #15208
• Implement Elasticsearch functions in GO by @qinling0210 in #15160
• Docs: Minimum required Python version increased to 3.13. by @writinwaters in #15219
• Fix: The prompt variable for the agent operator disappears after input. by @cike8899 in #15218
• fix: resolve asyncio correctness issues (fire-and-forget tasks, event loop nesting) by @wdeve...

v0.25.5

Summary

Released on May 20, 2026.

New features

• Adds local & SSH providers in admin panel. #15039

Improvements

• Accelerated dataset search path, reducing latency by 50–100% by removing expensive vector fetch and rerank similarity computation steps. #14970
• Pushes metadata filters down to Infinity, significantly speeding up metadata filtering. #14974
• added Redis caching for TTS. #14851
• atomic document counter updates #14867
• Improved server startup speed and memory usage #14973
• Agent: structured output aggregation #13384 #14848
• Agent: metadata filter reuse. #14849
• Optimizes connector dashboard. #14979

What's Changed

• fix: optimize reranking module robustness and bug fixes by @07heco in #14264
• fix: replace broken assert with raise ValueError in variable_assigner and loop by @Ricardo-M-L in #13906
• Fix: send input and output token usage to Langfuse by @Br1an67 in #13294
• Go: implement TTS for MiniMax provider and CLI testing for TTS by @Haruko386 in #14911
• fix: correct nested path traversal in set_variable_param_value by @Ricardo-M-L in #13986
• Fix WebDriver resource leak in HTML-to-PDF conversion by @Ricardo-M-L in #14310
• Doc: code component output section by @buua436 in #14915
• GO: align time units with Python and centralize timestamp injection in BaseModel by @buua436 in #14875
• Fix: missing authentication on agent file upload and download endpoints by @dale053 in #14854
• Go: fix LastLoginTime update by @buua436 in #14917
• fix: replace mutable default arguments with None in LLM chat models by @gambletan in #13513
• Go: update user settings fields by @buua436 in #14918
• fix: atomic chunk/token counter updates for documents and knowledge b… by @dale053 in #14867
• fix(agent): handle duplicate MCP tool names by @eviaaaaa in #14217
• fix: close db connections reliably in test_db_connection by @wdeveloper16 in #14777
• Go: implement TTS for fishaudio, openrouter and asr for fishaudio by @Haruko386 in #14926
• Feat: This enables SelectWithSearch to search by label. by @cike8899 in #14925
• Doc: Finalized v0.25.4 release notes by @writinwaters in #14929
• fix: add document download endpoint and refactor existing download function by @buua436 in #14927
• fix: preserve uploaded file attachments after subsequent assistant messages by @octo-patch in #13993
• fix(llm): Tongyi-Qianwen embeddings use correct DashScope native API for intl URLs by @sham-sr in #14784
• fix(agent): pass top_k and fix similarity weight slider behavior by @yingjianzh in #14760
• security: always use RestrictedUnpickler in deserialize_b64 (CWE-502) by @sebastiondev in #14803
• Go: add file parse command by @JinHai-CN in #14892
• fix(agentbot): aggregate structured output in non-streaming completions by @plind-junior in #14848
• fix: enable GitHub connector to sync PRs and issues by default by @octo-patch in #14062
• Go: implement Rerank in vLLM driver (#14878) by @hunnyboy1217 in #14880
• docs: add FAQ entry for using Ollama with RAGFlow by @SnakeEye-sudo in #14557
• Go: add cli command, list dataset documents by @JinHai-CN in #14948
• Go: implement TTS, ASR for Siliconflow and TTs for StepFun by @Haruko386 in #14944
• fix: correct attribute name typo model_speciess to model_species by @Ricardo-M-L in #13929
• feat: bump Python minimum from 3.12 to 3.13, drop strenum backport by @wdeveloper16 in #14767
• Go: implement PaddleOCR provider and implement ASR for CoHere by @Haruko386 in #14954
• Feat/web markdown UI updates by @xjq123456 in #14214
• fix: correct literal_eval dispatch and bool isinstance ordering in agent components by @Ricardo-M-L in #13988
• refactor: optimize BaseTitleChunker to improve RAG document chunk quality by @07heco in #14247
• Bump to infinity v0.7.0 by @qinling0210 in #14968
• Fix session deletion leaking chat-upload blobs by @6ba3i in #14969
• Go: implement Embed (embeddings) in Novita driver by @pandadev66 in #14895
• Go: add Jina chat completions support by @jakearmstrong59 in #14935
• Add Anthropic Go model provider by @carlos4s in #14940
• Fix: guard empty/whitespace embedding inputs in LLMBundle (#14428) by @Rene0422 in #14924
• Refactor: Drop the vector fetch for ES by @KevinHuSh in #14970
• Push metadata filters down to Infinity by @qinling0210 in #14974
• Go: implement provider: CometAPI by @tmimmanuel in #14930
• Fix: add SSRF guard for agent test_db_connection endpoint by @dale053 in #14860
• Go: fix siliconflow embedding response by @buua436 in #14975
• Refactor: speed up ragflow server, save startup memory by @wangq8 in #14973
• fix(api): enforce tenant access for connector routes by @dev111-actor in #14747
• Refactor: enhance graphrag - part 2 by @wangq8 in #14972
• Go: add restful api route aliases by @buua436 in #14977
• Go: implement provider: DeepInfra, XunFei by @Haruko386 in #14978
• fix: metadata_condition returning all docs when filter matches nothing by @hak2979 in #14967
• Bugfix: fix tag show by @wangq8 in #14980
• Fix admin CLI system variable commands by @jakearmstrong59 in #14956
• Go: fix forgetting policy validation and fix memory update diff checks by @buua436 in #14976
• Go: align document list response by @buua436 in #14982
• Go: add dataset graph api by @buua436 in #14984
• Feat: full optimization on connector dashboard by @Magicbook1108 in #14979
• Fix: validate memory tenant model IDs on update and enforce tenant scope in memory pipeline by @jony376 in #14923
• Misc: fix some typos by @kingloon in #14987
• Go: implement provider: MinerU by @Haruko386 in #14990
• Add Replicate chat provider by @tmimmanuel in #14958
• Fix: agent session log message by @buua436 in #14991
• Fix: MinerU vlm-http-client backend output file detection by @lksr1201 in #14240
• fix: the opencc c library uses fgets() to read dicti... in text.c by @orbisai0security in #13970
• feat(tts): cache synthesized speech in Redis to avoid redundant calls by @plind-junior in #14851
• fix(retrieval): keep manual metadata filter reusable inside Iteration by @plind-junior in #14849
• Go: implement provider: Xinference by @tmimmanuel in #14938
• Add TogetherAI chat provider by @tmimmanuel in #14957
• Fix: restore saved api_key fallback in add_llm (#14921) by @Rene0422 in #14941
• Feat: add new tests and tescases for restful api suite by @6ba3i in #14993
• Feat: VLM image descriptions in MinerU parser (#14869) by @...

v0.25.4

Summary

New features

• Adds a generic, configuration-driven RESTful API data source connector.

Improvements

• Agent tag management with filtering and sorting.
• Widget customization and persistence.

Model support

• Adds gpt-5.4-mini and gpt-5.4-nano to the OpenAI model list

Bug fixes

Fixed dataset document download route.

What's Changed

• Feature/generic api connector by @ahmadintisar in #13545
• fix: use context manager for pdfplumber to prevent resource leak by @gambletan in #13512
• Add widget customization and persistence by @47NoahThompson in #14603
• Docs: Editorial updates to the v0.25.3 release notes draft. by @writinwaters in #14903
• feat: add tag management for Agents with filtering and sorting (#14774) by @plind-junior in #14799
• Go: implement provider: Voyage AI by @tmimmanuel in #14811
• Fix doc format by @yingfeng in #14909
• Adds gpt-5.4-mini and gpt-5.4-nano by @sirj0k3r in #14908
• Docs: Finalized v0.25.3 release notes by @writinwaters in #14913
• Fix: dataset document download route by @buua436 in #14910
• Docs: Update version references to v0.25.4 in READMEs and docs by @asiroliu in #14912
• Docs: Added v0.25.4 release notes draft. by @writinwaters in #14914

New Contributors

• @47NoahThompson made their first contribution in #14603
• @plind-junior made their first contribution in #14799
• @sirj0k3r made their first contribution in #14908

Full Changelog: v0.25.3...v0.25.4

v0.25.3

Summary

New features

• Data source and parsing: Added column-level semantic/metadata control for the spreadsheet file parser; introduced ETag optimization for incremental synchronization of S3 data sources to avoid unnecessary file transfers.
• Enables assigning specific roles like content, metadata, and primary key, to table columns. #13710

Improvements

• API refactoring and security
  • Continues the transition of web APIs to RESTful conventions, ensuring backward compatibility for all legacy endpoints.
  • Binds the user_id in POST /api/v1/messages to the authenticated JWT principal. #14745
  • Secures the sandbox executor against dynamic and indirect code execution bypasses. #14690
  • LLM request timeout control
  • High concurrency blocking call thread pool
• Reduces ingestion server boot time. #14894

Bug fixes

• Images in multi-sheet Excel workbooks were not scoped by sheet, causing images to be incorrectly attributed across different worksheets. #14120
• Iteration item alias passing
• Tool parameter template parsing
• Code execution attachment output
• Volcano model addition fix

What's Changed

• Go: implement provider: Baidu by @Haruko386 in #14741
• feat(connectors): ETag-based bypass for incremental S3 ingestion (#14628) by @hunnyboy1217 in #14677
• Go: fix Baidu rerank issue by @JinHai-CN in #14742
• Go: fix siliconflow rerank issue by @JinHai-CN in #14743
• Go: implement Encode (embeddings) in OpenAI driver by @pandadev66 in #14630
• Fix: Radio.Group cloneElement crashes on non-element children by @JimZhang-lab in #14407
• fix(auth): fall back to session-based auth in _load_user by @mhtkarakose in #14569
• Fix: resolve template strings in tool component parameters by @wanghualoong in #14601
• fix base_url handling in HuggingfaceRerank by @Qwerrty574 in #14555
• Feature/table parser column roles by @ahmadintisar in #13710
• Feat: add BedrockCV for vision/image2text inference via LiteLLM by @vincentlambert in #14705
• Go: implement ListModels in Volcengine driver by @bittoby in #14702
• feat: make sandbox Dockerfile mirrors optional with ARG by @ParasSondhi in #14553
• fix(llm): add timeout to HTTP requests in LLM integration layer by @Ricardo-M-L in #14313
• Go: implement Encode (embeddings) in Google Gemini driver by @Joseff531 in #14682
• fix(go): wire Google CheckConnection to ListModels by @zeus1959 in #14660
• Fix(Go): correct Name() and region URL fallback in Aliyun driver by @Joseff531 in #14673
• fix: close two security analyzer bypass paths in sandbox executor by @Sp1kyss in #14690
• fix: handle missing parent chunk in retrieval_by_children by @vincentlambert in #14556
• Go: implement Encode (embeddings) in Gitee AI driver by @bittoby in #14698
• fix(keyword_extraction): accept Chinese commas/semicolons/newlines as keyword delimiters by @Qinsanz in #14540
• Go: implement Encode (embeddings) in vLLM driver by @pandadev66 in #14688
• fix: complete robustness fixes for rerank module addressing all review comments by @07heco in #14265
• fix(prompt): reserve system budget in message_fit_in by @hyl64 in #14164
• Go: implement Encode (embeddings) in Ollama driver by @jack-stormentswe in #14664
• Go: implement Encode (embeddings) in NVIDIA driver by @bittoby in #14700
• Perf(Go): batch SiliconFlow Encode requests with 32-item chunking by @Joseff531 in #14719
• Go: implement Encode (embeddings) in LM Studio driver by @pandadev66 in #14694
• Fix(Go): make OpenRouter Encode fail loudly on malformed responses by @Joseff531 in #14717
• Refactor: tidy up ThreadPoolExecutor lifecycle in file_service and task executor by @web-dev0521 in #14668
• GraphRAG feature - Part 1 - add spacy to extract entity and relation by @wangq8 in #14670
• fix: scope pending_cell_images by sheet in excel parser by @fplust in #14120
• fix(dify): guard retrieval argument error behavior by @Achieve3318 in #14169
• Fix: bind memory message user_id to authenticated user for JWT auth by @jony376 in #14745
• Fix: dataset search rerank id type by @buua436 in #14759
• Fix: shared dataset chunk index lookup by @buua436 in #14764
• fix: use context manager for ThreadPoolExecutor in file_service.py by @Ricardo-M-L in #14144
• Go: refactor embedding interface by @JinHai-CN in #14757
• Fix: safe argument guard and remove redundant redis call by @paulhuiseismic in #14060
• fix: offload blocking DB/Redis calls to thread pool for high-concurrency support (#13825) by @tmimmanuel in #13941
• Refact: Added a private helper _visibility_and_status_filter by @Sank-WoT in #13627
• Fix: Document parse status set to DONE before chunks are retrievable by @as-ondewo in #13352
• fix(web): fix incomplete Docx preview in citation reference by @yshchm in #14122
• fix: OCR.detect() returns truthy None-tuple causing NoneType subscript crash by @octo-patch in #13951
• chore: fix some comments to improve readability by @box4wangjing in #14756
• fix(opensearch): implement doc-meta dispatch surface on OSConnection by @tmimmanuel in #14577
• Go: add development guide document by @JinHai-CN in #14785
• Go: implement Rerank in NVIDIA driver by @RenzoMXD in #14778
• Fix: add codeexec attachments output by @buua436 in #14787
• Go: implement provider: CoHere and FishAudio by @Haruko386 in #14790
• Go: fix retrieval test error by @JinHai-CN in #14794
• [Bug]: REDIS error #12870 by @raminmardani in #13875
• fix(dify): add GET method support to /dify/retrieval for health check by @Lntanohuang in #13837
• feat(raptor): add Psi tree builder with original-space ranking and safe migration by @CaptainTimon in #14679
• Chore(deps): Bump urllib3 from 2.6.3 to 2.7.0 in /agent/sandbox by @dependabot[bot] in #14824
• Refactor(Go): remove hardcode in huggingface provider by @Haruko386 in #14822
• fix(agent): support iteration item aliases in child nodes by @hyl64 in #14146
• Go: implement provider: StepFun by @tmimmanuel in #14815
• fix(docs): correct broken knowledge graph construction link by @majiayu000 in #13838
• Fix: some agent bug by @buua436 in #14829
• Refact: sandbox quickstart.md & add tutorial for code exec component by @Magicbook1108 in #14786
• Test : aggregation edge cases for list and scalar values by @Achieve3318 in #14170
• Go: implement provider: Baichuan by @Haruko386 in #14832
• Go: implement Embed (embeddings) in Upstage driver by @tmimmanuel in #14819
• Speed up start time by @wangq8 in #14833
• GO: implement GET /api/v1/datasets/:dataset_id by @buua436 in #14834
• Go: add ASR, TTS, OCR command by @JinHai-CN in #14836
• Go: fix dataset time unit by @buua436 in #14837
• Go: implement Embed (embeddings) in Mistral driver by @tmimmanuel in #14807
• Go: implement provider: Jina by @Haruko386 in #14838
  ...