Error Details: Custom error: text.replace is not a function

miguelservent · August 26, 2025, 5:49pm

Hi, I am getting an error while making simple edits of a simple project. Please, advise what can be done here to fix it.

rajputanakshay1 · September 10, 2025, 11:49am

@xKevIsDev seems like sanitizeText function written in streamtext is causing an issue

rajputanakshay1 · September 17, 2025, 6:05am

@xKevIsDev
We are receiving an array as the data type for text instead of a string which is causing the error. This issue occurs only when clicking the buttons received in response in Discuss mode. Since you originally wrote the code for Discuss mode, I think you should be able to fix it quickly.

private.winters.bf3 · September 17, 2025, 8:13pm

Hi,

Can you try this PR :

github.com/stackblitz-labs/bolt.diy

feat: comprehensive major SDK update with advanced LLM and service integrations

main ← Stijnus:BOLTDIY_MAJOR_UPDATE_SDK

opened 10:22AM - 16 Sep 25 UTC

Stijnus

+5231 -2240

## 🚀 Major SDK Update - Comprehensive Feature Enhancement This PR represents a …**massive enhancement** to bolt.diy with **241 files changed** (30,663 additions, 10,916 deletions), introducing advanced AI capabilities, comprehensive service integrations, and significant architectural improvements. ### 🎯 **Core Features** #### 🤖 **Advanced LLM Provider System** - **New Providers**: Cerebras (ultra-fast inference), Cloudflare Workers AI, Moonshot AI (Kimi) - **Enhanced Providers**: Comprehensive updates to OpenAI, Anthropic, Google, Groq, DeepSeek, xAI, Perplexity, and more - **Provider-Specific Prompt Optimization**: Intelligent prompt optimization based on provider capabilities - 60% token reduction for speed-optimized providers (Groq) - 40% reduction for reasoning models (o1) - let them reason internally - 45% reduction for local models (Ollama) - resource efficiency - 20% expansion for high-context models (Google) - utilize full capabilities - **Dynamic Token Management**: 4-tier capability detection with automatic context optimization - **Reasoning Model Support**: Special handling for o1, o3 with parameter filtering #### 🔌 **Comprehensive Service Integration Refactor** - **GitHub Integration**: Complete overhaul with repository management, branch selection, stats, deployment - **GitLab Integration**: Full GitLab support with project management, deployment capabilities - **Enhanced Netlify**: Advanced deployment features with user management - **Vercel Integration**: Comprehensive deployment and project management - **Supabase Integration**: Database project management with query capabilities - **Unified Service Architecture**: Consistent patterns across all service integrations #### 🧠 **Enhanced Message Parser & Runtime** - **AI Model Output Pattern Integration**: Handles GPT-4, Claude, Gemini output patterns - **Auto-Artifact Generation**: Automatically wraps code blocks when structured tags missing - **Enhanced File Change Optimizer**: Runtime optimization for better performance during streaming - **Stream Recovery System**: Prevents chat hanging with robust error handling - **Smart Shell Command Detection**: Intelligent detection of npm, git, docker commands #### 🏗️ **Advanced Architecture Improvements** - **Service Tabs Refactor**: Modern tabbed architecture for better UX - **Health Monitoring**: Local model health monitoring with real-time status - **Debug & Logging System**: Comprehensive debugging with capture and download capabilities - **Security Enhancements**: CODEOWNERS, advanced GitHub workflows, security scanning - **Storage Utilities**: Advanced storage management with caching strategies ### 📊 **Technical Highlights** #### 🎛️ **Provider Categories & Optimization** - **High-Context**: Google Gemini (expanded prompts, detailed guidelines) - **Speed-Optimized**: Groq, Cerebras (ultra-concise prompts, minimal latency) - **Reasoning**: OpenAI o1/o3 (simplified instructions, let model reason) - **Local Models**: Ollama (resource-efficient, simplified language) - **Coding-Specialized**: DeepSeek, GitHub Copilot (enhanced code guidelines) - **Standard**: OpenAI GPT-4o (balanced approach) #### ⚡ **Performance Optimizations** - **Context Optimization**: AI-powered file selection and chat summarization - **Token Management**: Dynamic context window utilization per provider - **Streaming Improvements**: Real-time parsing with abort capabilities - **File Operations**: Optimized file change detection and processing #### 🛡️ **Security & Quality** - **GitHub Workflows**: Preview, quality, security scanning, test automation - **Code Quality**: Enhanced ESLint rules, consistent formatting - **Type Safety**: Comprehensive TypeScript improvements - **Testing**: Enhanced test coverage with real-world patterns ### 🎨 **UI/UX Enhancements** - **Enhanced Tabs**: Modern service integration interface - **Repository Management**: Advanced GitHub/GitLab repository selection - **Model Selector**: Improved provider and model selection UI - **Status Indicators**: Real-time connection and health status - **Error Handling**: Better error states and user feedback - **Responsive Design**: Improved mobile and desktop layouts ### 🔧 **Developer Experience** - **Modular Architecture**: Clean separation of concerns - **Provider Registry**: Easy addition of new LLM providers - **Hook System**: Comprehensive React hooks for service integrations - **Type Definitions**: Enhanced TypeScript definitions across all services - **Documentation**: Comprehensive inline documentation ### 📈 **Impact & Benefits** - **Faster Inference**: Up to 60% token reduction for speed-optimized providers - **Better Quality**: Provider-specific optimizations improve response quality - **Resource Efficiency**: Optimized for local models and limited resources - **Enhanced Capabilities**: Full utilization of high-context models - **Improved Reliability**: Robust error handling and recovery systems - **Better UX**: Streamlined service integrations and modern UI ### 🧪 **Testing & Quality Assurance** - ✅ **Build**: Passes production build - ✅ **Linting**: No ESLint issues - ✅ **TypeCheck**: No TypeScript errors - ✅ **Tests**: All test suites passing - ✅ **Pre-commit Hooks**: All quality checks pass ### 📋 **Migration Notes** - **Backward Compatible**: Existing configurations and prompts continue to work - **Service Connections**: Existing service connections are preserved - **Model Configurations**: All existing model configurations are maintained - **API Compatibility**: No breaking changes to existing APIs ### 🎉 **Ready for Production** This comprehensive update represents months of development work, bringing bolt.diy to the next level with: - **20+ LLM Providers** with intelligent optimization - **5+ Service Integrations** with advanced features - **Advanced AI Capabilities** with provider-specific intelligence - **Modern Architecture** with enhanced performance and reliability 🤖 Generated with [Claude Code](https://claude.ai/code)

Br,

Stijnus