ragflow Releases

A feature-packed update introducing mobile knowledge packs, GPU acceleration for chat responses on Linux, and performance optimizations like KV caching and improved download management.

Aug 5, 2026

This release enables Cap to function end-to-end with AI agents via CLI and MCP servers, eliminating the need for the dashboard. It enhances the recording experience with a built-in teleprompter, clip transitions, and privacy masks, while also improving editor performance.

Jul 20, 2026

This release addresses critical stability issues regarding recording behavior on high-refresh monitors and multi-monitor setups. It adds user-requested features like microphone muting during recording and a nightly update channel to the application.

Jul 17, 2026

Focuses on NLP improvements with a Snowball stemmer supporting 16 languages and Dutch support. Includes significant bug fixes for LM-Studio crashes, MCP server errors, Docling parsers, and metadata persistence.

Jul 7, 2026

Introduces Google BigQuery as a data source connector, adds MCP tools for listing datasets and chats, and integrates the SoMark layout-aware OCR parser. Also adds partial success handling for batch uploads.

Jul 2, 2026

This release introduces screenshot sharing capabilities, expands desktop automation support with CLI integration, and enhances the editing suite with audio mixing and advanced captioning tools. It also significantly improves Linux compatibility and addresses reliability issues across uploads, recordings, and previews.

Jul 1, 2026

Jun 30, 2026

Jun 29, 2026

Adds new chat channels for WhatsApp, DingTalk, and WeCom. Improves PaddleOCR with fallback logic and integrates image parsing. Fixes various parsing bugs in DOCX and Laws parsers, as well as MCP and dataflow issues.

Jun 29, 2026

Focusing on performance and automation, this release adds robust desktop automation rules, improves the Cap CLI with installation and repair utilities, and introduces split-screen scene composition. It also refines caption editing and fixes various authentication and platform-specific stability issues.

Jun 27, 2026

Jun 22, 2026

Allows users to modify model types in the model provider, enables Discord and Feishu chat channels, and groups multi-turn chat traces by session in Langfuse. Fixes token billing and adds a legacy mode to the chat completions API.

Jun 17, 2026

Major release implementing auto-populated model lists for multiple providers, new data source connectors (Outlook, OneDrive, Teams, Slack, SharePoint, Salesforce, Azure Blob), and a checkpoint/resume feature for GraphRAG indexing.

Jun 11, 2026

This release focuses on security hardening, raising minimum supported versions for Deno and Node, and patching multiple CVEs. It introduces new extractor capabilities, lockfile management, and updates to the GitHub REST API.

Jun 9, 2026

Cap Desktop 0.5.1 enhances the recording experience for both human users and automated agents by bundling the Cap CLI and introducing advanced editing features like split-screen mode. The release also improves recovery mechanisms for fragmented uploads, audio timing, and macOS stability.

Jun 8, 2026

Adds a Browser component for AI web navigation, introduces AHC mode (Ψ-RAG) for RAPTOR, and adds a lightweight `@tool` decorator. Fixes various bugs related to streaming responses, parsing, and agent components.

May 27, 2026

Accelerates dataset search path by 50-100%, pushes metadata filters down to Infinity, and adds Redis caching for TTS. Introduces local & SSH providers in the admin panel and improves connector dashboard.

May 20, 2026

Introduces a generic RESTful API data source connector, adds agent tag management with filtering and sorting, and enables widget customization and persistence. Adds gpt-5.4-mini and gpt-5.4-nano to the model list.

May 14, 2026

Continues the transition to RESTful APIs and enhances security, introduces spreadsheet column-level control and ETag optimization for S3 data sources, and reduces ingestion server boot time.

May 13, 2026

This is a small patch release focused on bug fixes for DeepSeek V4, gpt-oss, and Qwen3-VL models, addressing issues with sparse attention, KV cache allocation, and deepstack boundaries.

May 10, 2026