293 lines (196 loc) · 14.2 KB

English | 简体中文

Chats Release History

This page indexes all major version release notes for the Chats project, from the latest to earlier versions. Each version includes core features, improvements, and fixes.

1.10.2 - 2026-03-10 ⭐ Latest Release

Core Highlights: Request Trace end-to-end tracing · Admin request trace workspace · Sampling/filtering/redaction/retention policies · Export, compare, and scheduled cleanup · Small UX and compatibility improvements

🔍 Request Trace: adds inbound and outbound HTTP request tracing with timing points, status codes, errors, body capture, and raw payload downloads
🧑‍💼 Admin Workspace: adds a dedicated request trace page with filters, 24-column customization, details, raw downloads, two-trace compare, Excel export, and bulk delete by filters
⚙️ Trace Governance: supports separate inbound/outbound settings, sample rate, include/exclude filters, URL parameter redaction, and retention policies
♻️ Lifecycle Management: adds bounded queue protection and scheduled cleanup to reduce long-running trace storage pressure
🛠️ Other Improvements: updates dashboard summary card metrics, adds long user message collapse, and improves Anthropic protocol compatibility

View Full Release Notes →

1.10.1 - 2026-02-15

Core Highlights: Sandbox Manager Enhancements · Server-side ETag Caching · ChatMiniMap Navigation · API Key Management Upgrades · Code Interpreter Image Pipeline Updates

🏷️ Sandbox Manager: Session Manager renamed and expanded with info/env/files/editor capabilities
⚡ ETag Caching: High-frequency APIs now support server-driven ETag + 304 responses, frontend chat list local cache removed
🗺️ ChatMiniMap: Adds right-side message mini-map and migrates scroll controls from ChatInput
🔐 API Keys: One-time full-key display on create, masked list responses, edit + bulk copy support
🐳 Image Pipeline: RUN_NUMBER rv marker in skills, ripgrep preinstalled, and non-main branches can publish latest manifest

View Full Release Notes →

1.10.0 - 2026-01-14

Core Highlights: Built-in Docker Code Interpreter · sdcb/code-interpreter Image & CI · Multi-provider Interleaved Thinking · DeepSeek v3.2 reasoning_content Compatibility · Tooling/Upload/Admin UI Improvements

🐳 Built-in Code Interpreter: New Docker-based toolset with session management (create/run/read/write/patch/download/destroy)
📦 Default Image: sdcb/code-interpreter as the out-of-box environment (multi-arch), preloaded with toolchains, document/media utilities, browser stack, and anthropics/skills
🧠 Interleaved Thinking: Moonshot/DeepSeek/MiMo support reasoning (reasoning_content) round-tripping across tool calls
🔁 DeepSeek v3.2 Compatibility: Chat Completions API parses/persists assistant reasoning_content for stable think–tool–think flows
🧰 Frontend UX: ToolCallBlock progress/long output display, drag/paste upload, file preview & share UI, generation info & dashboard improvements

View Full Release Notes →

1.9.1 - 2025-12-21

Core Highlights: Xiaomi MiMo Provider · Prompt Cache Token Billing · Next.js 16 & React 19 Upgrade · Interleaved Thinking · 4-Level Chat View Architecture

🤖 Xiaomi MiMo Support: New Xiaomi MiMo provider (ID=21), supports MiMo-V2-Flash with OpenAI/Anthropic API formats
💰 Prompt Cache Billing: Cache token pricing, distinguishes Fresh/Cached tokens, displays usage in generation info
🔄 Framework Upgrades: Next.js 15.5.3 → 16.0.7, React 18.2.0 → 19.2.1, upgraded all @radix-ui packages
🧠 Interleaved Thinking: Minimax and DeepSeek support for Chat Completions API format interleaved thinking
🏗️ 4-Level Architecture: chat/turn/step/content structure, generation info bubbles, improved message display
📋 LaTeX Copy: One-click copy of original LaTeX formula code with custom rehype plugin
🗑️ SDK Cleanup: Removed OpenAI .NET SDK and Mscc.GenerativeAI, using native HttpClient
🐛 Bug Fixes: DeepSeek-R1 signature parsing, token counting, image generation, UI flickering

View Full Release Notes →

1.9.0 - 2025-11-27

Core Highlights: Anthropic Provider Support · Anthropic Messages API · OpenAI Image API · Build Developer Pages · .NET 10 Upgrade

🤖 Anthropic Support: Full support for Claude model series (ID=20), including thinking+signature flow
📡 Anthropic Messages API: Compatible with Anthropic API spec (/v1/messages)
🖼️ OpenAI Image API: Image generation (/v1/images/generations) and edit (/v1/images/edits)
🛠️ Build Pages: Developer-facing API Keys / Docs / Usage management pages
🎬 UI Animations: ChatInput expand/fullscreen, UserMenuPopover, ToolCallBlock animations
⬆️ .NET 10 Upgrade: Framework upgraded to .NET 10
🏗️ Architecture Refactor: ChatService changed to DB Steps driven, supporting multiple message formats
🐛 Bug Fixes: Thinking Budget, Gemini Thinking, UI flashing issues

View Full Release Notes →

1.8.1 - 2025-11-11

Core Highlights: User Model Permission Management · Reasoning Lifecycle Tracking · File Preview Refactor · Chat Cache Optimization

👥 User Model Permission Management: New permission management system with dual perspective (by user and by model)
🧠 Reasoning Lifecycle Tracking: Auto-track reasoning state, smart expand/collapse reasoning content
📎 File Preview Refactor: Unified file preview component supporting images, videos, audio, documents, and more
⚡ Chat List Caching: localStorage cache + parallel loading, significantly improved first-screen load speed
🐴 TokenPony Provider: Added TokenPony provider support (ID=19)
🔄 Azure Branding Upgrade: Azure OpenAI → Azure AI Foundry
🛠️ ChatService Simplification: Architecture simplified, removed redundant code
🐛 OpenAI 2.6.0 Fix: Switched to self-compiled Sdcb.OpenAI package, fixed reasoning content parsing issue

View Full Release Notes →

1.8.0 - 2025-11-04 🎉 Major Architecture Upgrade

Core Highlights: Model Configuration Architecture Refactor · User-Customizable Config · No Migration for New Models

🏗️ Architecture Refactor: Model config migrated from ModelReference static table to Model instance level
🎯 User Customization: Each model instance has independent complete configuration (18 new fields)
🚀 Ready to Use: Adding new models requires no database migration, users can freely configure
🗑️ Simplified Architecture: Deleted 6 static reference tables (ModelReference, ModelProvider, etc.)
🖼️ Third API Type: Added ImageGeneration type (ChatCompletion, Response, ImageGeneration)
🔍 Model Validation Framework: Custom validation attributes + 436 lines of unit tests
⚡ Quick Add Models: New batch add dialog, supports continuous additions
📸 Full-screen Image Preview: Zoom animation + keyboard navigation + thumbnail strip
🎨 UI Optimizations: Admin on-demand lazy loading, skeleton screens, separated preset configs
🛠️ Image Processing Refactor: ImageSharp replaces hand-written parsers (-1,217 lines)

View Full Release Notes → | API Changes

1.7.2 - 2025-10-27

Core Highlights: Streaming Image Generation · Code Execution · API Test Framework

🖼️ Streaming Image Generation: Real-time progress for gpt-image-1 and gpt-image-1-mini
🔧 Code Execution: Gemini code execution results displayed as tool calls
🧪 API Test Framework: New Chats.BE.ApiTest project for comprehensive OpenAI compatible API testing
🤖 New Models: Added gpt-5-codex, gpt-5-pro, gpt-image-1-mini
⚙️ Config Enhancement: New CodeExecutionEnabled field in ChatConfig
📦 Dependency Updates: Upgraded 11 third-party packages

View Full Release Notes →

1.7.1 - 2025-10-13

Core Highlights: Security Audit Logs · Send Experience · Tool Call Display

🔒 Security Audit: Complete login attempt records and rate limit monitoring
📤 Send Experience: Multiple send modes (send, continue, regenerate) with mobile adaptation
🔨 Tool Calls: Dedicated ToolCallBlock component for clearer parameter and result display
⚡ Performance: On-demand loading of generation info, reduced initial load time
📱 Mobile: Better responsive layout and touch experience
👨‍💼 Admin Panel: Message content queries, optimized user usage statistics

View Full Release Notes →

1.7.0 - 2025-09-20 🎉 Major Update

Core Highlights: Full MCP Support · Database Refactor · Drag-and-Drop Ordering

🔌 MCP Support: End-to-end server/frontend integration, user authorization, tool discovery
🛠️ Enhanced Tool Calls: Richer SSE events, new tool request/response message content types
🗄️ Database Refactor: Message→ChatTurn/Step layering, data migration (breaking change)
🎨 Drag-and-Drop: Support for models/keys/presets ordering
📊 Mermaid Support: New Markdown Mermaid renderer with dark/light themes
🖼️ Image Sizes: Common size options (1024×1024, 1536×1024, 1024×1536)
🔄 Regeneration: Support for single message or whole segment regeneration

View Full Release Notes →

1.6 - 2024-06-30

Core Highlights: .NET 9 Upgrade · Enhanced Reasoning Models · Image Improvements

⬆️ .NET 9: Full upgrade to .NET 9 framework for improved performance and security
🧠 Reasoning Models: Support for o3-pro, o4-mini, Gemini Think
📸 Image Features: History upload, phone camera capture, forced 3:2/2:3 ratios
🤖 New Models: GPT-5 series, Qwen3 series, GLM-4.5 series, Kimi, Grok-4, Codex-mini
☁️ Azure Deployment: New one-click deployment script
💾 Storage: Migrated from MinIO to Cloudflare R2
⌨️ UX: Fullscreen hotkey (Ctrl+F), UI optimizations

View Full Release Notes →

1.5 - 2024-05-20

Core Highlights: Qwen3 Support · Stability Improvements

🤖 Qwen3 Support: Added Qwen3 models with optional reasoning chain disable
🗂️ File Optimization: Database indexes for better performance, improved download logic
🎨 UI Enhancement: Optimized dark theme icons, improved code/raw content toggle
🐛 Bug Fixes: Image generation, file sharing, encryption/decryption fixes
📦 Dependency Updates: Upgraded to latest package versions

View Full Release Notes →

1.4.0 - 2024-05-20

Core Highlights: Full Function Call Support · API Caching

🔧 Function Calls: All API endpoints support function calling
⚡ API Caching: New /v1-cached and /v1-cached-createOnly cache endpoints
💻 Code Execution: Google Gemini models support code execution
🚀 Performance: Async processing, reduced duplicate database calls

View Full Release Notes →

1.3 - 2024-04-25

Core Highlights: Enhanced Image Generation · Baidu ERNIE Integration

🖼️ gpt-image-1: Full support for Azure OpenAI gpt-image-1 model
✏️ Image Editing: Image editing, mask redrawing, quality control
🤖 Baidu ERNIE: Integration of new Baidu ERNIE models
📁 File Management: Improved file management features
🐛 Bug Fixes: 1.3.1 fixed critical password change issue

View Full Release Notes →

1.2.0 - 2025-04-25

Core Highlights: Admin Dashboard · Data Visualization

📊 Dashboard: Data visualization panel for chat volume, cost, and token statistics
📱 Mobile Adaptation: Full mobile device support for admin panel
💬 Message UI: Improved message display styles, new message type toggle
⌨️ Interaction: Message edit with Ctrl+Enter, search debounce optimization

View Full Release Notes →

1.1.0 - 2025-04-24

Core Highlights: Reasoning Model Support · SDK Upgrade

🧠 Reasoning Models: Azure OpenAI Response API with o3/o4-mini reasoning summary
📦 SDK Upgrade: OpenAI and Azure OpenAI SDK upgraded to 2.2.0-beta.4
📝 Reasoning Optimization: Improved reasoning process format for readability
🌏 Mirror Upload: Sync upload to Minio for faster China access

View Full Release Notes →

1.0 - 2025-04-21 🎊 Official Release

Core Highlights: Official Release · Comprehensive Features

🎉 Official Release: From 0.x preview to production-ready
⚙️ User Settings: New user settings page for centralized personal configuration
📈 Usage Reports: Detailed request and consumption tracking with Excel export
🤖 Model Expansion: GPT-4.1, o3/o4-mini, Doubao 1.5, Gemini 2.5 Flash
⏱️ Timeout Optimization: Thinking timeout extended from 100s to 24 hours
🔄 Background Generation: Model continues generating after window close
👨‍💼 Admin Features: Consumption summary with multi-condition filtering

Version Notes:

1.0.0 (756): Official release

1.0.1 (759): Fixed reasoning level issues

1.0.2 (762): Fixed Google Gemini reasoning level issues

View Full Release Notes →

Version Naming Convention

Starting from 1.0.0, Chats follows Semantic Versioning:

Major: Significant architectural changes or breaking updates (e.g., 1.0.0 → 2.0.0)
Minor: New features with backward compatibility (e.g., 1.0.0 → 1.1.0)
Patch: Bug fixes and minor optimizations (e.g., 1.0.0 → 1.0.1)

Get Help

_{Last updated: 2025-12-21}