* fix: improve local model provider robustness and UX
- Extract shared Docker URL rewriting and env conversion into BaseProvider
to eliminate 4x duplicated code across Ollama and LMStudio
- Add error handling and 5s timeouts to all model-listing fetches so one
unreachable provider doesn't block the entire model list
- Fix Ollama using createOllama() instead of mutating provider internals
- Fix LLMManager singleton ignoring env updates on subsequent requests
- Narrow cache key to only include provider-relevant env vars instead of
the entire server environment
- Fix 'as any' casts in LMStudio and OpenAILike by using shared
convertEnvToRecord helper
- Replace console.log/error with structured logger in OpenAILike
- Fix typo: filteredStaticModesl -> filteredStaticModels in manager
- Add connection status indicator (green/red dot) for local providers
in the ModelSelector dropdown
- Show helpful "is X running?" message when local provider has no models
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* feat: add Cerebras LLM provider
- Add Cerebras provider with 8 models (Llama, GPT OSS, Qwen, ZAI GLM)
- Integrate @ai-sdk/cerebras@0.2.16 for compatibility
- Add CEREBRAS_API_KEY to environment configuration
- Register provider in LLMManager registry
Models included:
- llama3.1-8b, llama-3.3-70b
- gpt-oss-120b (reasoning)
- qwen-3-32b, qwen-3-235b variants
- zai-glm-4.6, zai-glm-4.7 (reasoning)
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* feat: add Fireworks AI LLM provider
- Add Fireworks provider with 6 popular models
- Integrate @ai-sdk/fireworks@0.2.16 for compatibility
- Add FIREWORKS_API_KEY to environment configuration
- Register provider in LLMManager registry
Models included:
- Llama 3.1 variants (405B, 70B, 8B Instruct)
- DeepSeek R1 (reasoning model)
- Qwen 2.5 72B Instruct
- FireFunction V2
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* feat: add coding-specific models to existing providers
Enhanced providers with state-of-the-art coding models:
**DeepSeek Provider:**
+ DeepSeek V3.2 (integrates thinking + tool-use)
+ DeepSeek V3.2-Speciale (high-compute variant, beats GPT-5)
**Fireworks Provider:**
+ Qwen3-Coder 480B (262K context, best for coding)
+ Qwen3-Coder 30B (fast coding specialist)
**Cerebras Provider:**
+ Qwen3-Coder 480B (2000 tokens/sec!)
- Removed deprecated models (qwen-3-32b, llama-3.3-70b)
Total new models: 4
Total coding models across all providers: 12+
Performance highlights:
- Qwen3-Coder: State-of-the-art coding performance
- DeepSeek V3.2: Integrates thinking directly into tool-use
- ZAI GLM 4.6: 73.8% SWE-bench score
- Ultra-fast inference: 2000 tok/s on Cerebras
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* feat: add dynamic model discovery to providers
Implemented getDynamicModels() for automatic model discovery:
**DeepSeek Provider:**
- Fetches models from https://api.deepseek.com/models
- Automatically discovers new models as DeepSeek adds them
- Filters out static models to avoid duplicates
**Cerebras Provider:**
- Fetches models from https://api.cerebras.ai/v1/models
- Auto-discovers new Cerebras models
- Keeps UI up-to-date with latest offerings
**Fireworks Provider:**
- Fetches from https://api.fireworks.ai/v1/accounts/fireworks/models
- Includes context_length from API response
- Discovers new Qwen-Coder and other models automatically
**Moonshot Provider:**
- Fetches from https://api.moonshot.ai/v1/models
- OpenAI-compatible endpoint
- Auto-discovers new Kimi models
Benefits:
- ✅ No manual updates needed when providers add new models
- ✅ Users always have access to latest models
- ✅ Graceful fallback to static models if API fails
- ✅ 5-second timeout prevents hanging
- ✅ Caching system built into BaseProvider
Technical details:
- Uses BaseProvider's built-in caching system
- Cache invalidates when API keys change
- Failed API calls fallback to static models
- All endpoints have 5-second timeout protection
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* feat: add Z.AI provider with GLM models and JWT authentication
Merged changes from PR #2069 to add Z.AI provider:
- Added GLM-4.6 (200K), GLM-4.5 (128K), and GLM-4.5 Flash models
- Implemented secure JWT token generation with HMAC-SHA256 signing
- Added dynamic model discovery from Z.AI API
- Included proper error handling and token validation
- GLM-4.6 achieves 73.8% on SWE-bench coding benchmarks
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
* feat: add service tabs refactor with GitHub, GitLab, Supabase, Vercel, and Netlify integration
This commit introduces a comprehensive refactor of the connections system,
replacing the single connections tab with dedicated service integration tabs:
✨ New Service Tabs:
- GitHub Tab: Complete integration with repository management, stats, and API
- GitLab Tab: GitLab project integration and management
- Supabase Tab: Database project management with comprehensive analytics
- Vercel Tab: Project deployment management and monitoring
- Netlify Tab: Site deployment and build management
🔧 Supporting Infrastructure:
- Enhanced store management for each service with auto-connect via env vars
- API routes for secure server-side token handling and data fetching
- Updated TypeScript types with missing properties and interfaces
- Comprehensive hooks for service connections and state management
- Security utilities for API endpoint validation
🎨 UI/UX Improvements:
- Individual service tabs with tailored functionality
- Motion animations and improved loading states
- Connection testing and health monitoring
- Advanced analytics dashboards for each service
- Consistent design patterns across all service tabs
🛠️ Technical Changes:
- Removed legacy connection tab in favor of individual service tabs
- Updated tab configuration and routing system
- Added comprehensive error handling and loading states
- Enhanced type safety with extended interfaces
- Implemented environment variable auto-connection features
Note: Some TypeScript errors remain and will need to be resolved in follow-up commits.
The dev server runs successfully and the service tabs are functional.
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
* feat: comprehensive service integration refactor with enhanced tabs architecture
Major architectural improvements to service integrations:
**Service Integration Refactor:**
- Complete restructure of service connection tabs (GitHub, GitLab, Vercel, Netlify, Supabase)
- Migrated from centralized ConnectionsTab to dedicated service-specific tabs
- Added shared service integration components for consistent UX
- Implemented auto-connection feature using environment variables
**New Components & Architecture:**
- ServiceIntegrationLayout for consistent service tab structure
- ConnectionStatus, ServiceCard components for reusable UI patterns
- BranchSelector component for repository branch management
- Enhanced authentication dialogs with improved error handling
**API & Backend Enhancements:**
- New API endpoints: github-branches, gitlab-branches, gitlab-projects, vercel-user
- Enhanced GitLab API service with comprehensive project management
- Improved connection testing hooks (useConnectionTest)
- Better error handling and rate limiting across all services
**Configuration & Environment:**
- Updated .env.example with comprehensive service integration guides
- Added auto-connection support for all major services
- Improved development and production environment configurations
- Enhanced tab management with proper service icons
**Code Quality & TypeScript:**
- Fixed all TypeScript errors across service integration components
- Enhanced type definitions for Vercel, Supabase, and other service integrations
- Improved type safety with proper optional chaining and type assertions
- Better separation of concerns between UI and business logic
**Removed Legacy Code:**
- Removed redundant connection components and consolidated into service tabs
- Cleaned up unused imports and deprecated connection patterns
- Streamlined authentication flows across all services
This refactor provides a more maintainable, scalable architecture for service integrations
while significantly improving the user experience for managing external connections.
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
* refactor: clean up dead code and consolidate utilities
- Remove legacy .eslintrc.json (replaced by flat config)
- Remove duplicate app/utils/types.ts (unused type definitions)
- Remove app/utils/cn.ts and consolidate with classNames utility
- Clean up unused ServiceErrorHandler class implementation
- Enhance classNames utility to support boolean values
- Update GlowingEffect.tsx to use consolidated classNames utility
Removes ~150+ lines of unused code while maintaining all functionality.
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
* Simplify terminal health checks and improve project setup
Removed aggressive health checking and reconnection logic from TerminalManager to prevent issues with terminal responsiveness. Updated TerminalTabs to remove onReconnect handlers. Enhanced projectCommands utility to generate non-interactive setup commands and detect shadcn projects, improving automation and reliability of project setup.
* fix: resolve GitLab deployment issues and enhance GitHub deployment reliability
GitLab Deployment Fixes:
- Fix COEP header issue for avatar images by adding crossOrigin and referrerPolicy attributes
- Implement repository name sanitization to handle special characters and ensure GitLab compliance
- Enhance error handling with detailed validation error parsing and user-friendly messages
- Add explicit path field and description to project creation requests
- Improve URL encoding and project path resolution for proper API calls
- Add graceful file commit handling with timeout and error recovery
GitHub Deployment Enhancements:
- Add comprehensive repository name validation and sanitization
- Implement real-time feedback for invalid characters in repository name input
- Enhance error handling with specific error types and retry suggestions
- Improve user experience with better error messages and validation feedback
- Add repository name length limits and character restrictions
- Show sanitized name preview to users before submission
General Improvements:
- Add GitLabAuthDialog component for improved authentication flow
- Enhance logging and debugging capabilities for deployment operations
- Improve accessibility with proper dialog titles and descriptions
- Add better user notifications for name sanitization and validation issues
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
---------
Co-authored-by: Claude <noreply@anthropic.com>
- Add OPENAI_LIKE_API_MODELS environment variable support
- Enable fallback model parsing when /models endpoint fails
- Support providers like Fireworks AI that don't allow /models requests
- Format: path/to/model1:limit;path/to/model2:limit;path/to/model3:limit
- Update IProviderSetting interface to include OPENAI_LIKE_API_MODELS property
- Fix all linting errors and code formatting issues
- Add comprehensive Moonshot AI provider with 11 models including:
* Legacy moonshot-v1 series (8k, 32k, 128k context)
* Latest Kimi K2 models (K2 Preview, Turbo, Thinking)
* Vision-enabled models for multimodal capabilities
* Auto-selecting model variants
- Update xAI provider with latest Grok models:
* Add Grok 4 (256K context) and Grok 4 (07-09) variant
* Add Grok 3 Mini Beta and Mini Fast Beta variants
* Update context limits to match actual model capabilities
* Remove outdated grok-beta and grok-2-1212 models
- Add MOONSHOT_API_KEY to environment configuration
- Register Moonshot provider in service status monitoring
- Full OpenAI-compatible API integration via api.moonshot.ai
- Fix TypeScript errors in GitHub provider
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-authored-by: Claude <noreply@anthropic.com>
* Update LLM providers and constants
- Updated constants in app/lib/.server/llm/constants.ts
- Modified stream-text functionality in app/lib/.server/llm/stream-text.ts
- Updated Anthropic provider in app/lib/modules/llm/providers/anthropic.ts
- Modified GitHub provider in app/lib/modules/llm/providers/github.ts
- Updated Google provider in app/lib/modules/llm/providers/google.ts
- Modified OpenAI provider in app/lib/modules/llm/providers/openai.ts
- Updated LLM types in app/lib/modules/llm/types.ts
- Modified API route in app/routes/api.llmcall.ts
* Fix maxCompletionTokens Implementation for All Providers
- Cohere: Added maxCompletionTokens: 4000 to all 10 static models
- DeepSeek: Added maxCompletionTokens: 8192 to all 3 static models
- Groq: Added maxCompletionTokens: 8192 to both static models
- Mistral: Added maxCompletionTokens: 8192 to all 9 static models
- Together: Added maxCompletionTokens: 8192 to both static models
- Groq: Fixed getDynamicModels to include maxCompletionTokens: 8192
- Together: Fixed getDynamicModels to include maxCompletionTokens: 8192
- OpenAI: Fixed getDynamicModels with proper logic for reasoning models (o1: 16384, o1-mini: 8192) and standard models
- Updated constants in app/lib/.server/llm/constants.ts
- Modified stream-text functionality in app/lib/.server/llm/stream-text.ts
- Updated Anthropic provider in app/lib/modules/llm/providers/anthropic.ts
- Modified GitHub provider in app/lib/modules/llm/providers/github.ts
- Updated Google provider in app/lib/modules/llm/providers/google.ts
- Modified OpenAI provider in app/lib/modules/llm/providers/openai.ts
- Updated LLM types in app/lib/modules/llm/types.ts
- Modified API route in app/routes/api.llmcall.ts
- Changed the maxTokenAllowed property to use Math.min for limiting the value to a maximum of 16384 tokens, ensuring better control over context window size.
- Added logging for dynamic max tokens based on model details.
- Increased max token limit for Claude model from 8000 to 128000.
- Included beta header for Anthropik API call.
Add 'grok-3-beta' to xAI provider and 'gemini-2.5-flash-preview-04-17' to Google provider. Also, ensure file saving when content is updated in WorkbenchStore and update streaming indicator styling in chat messages.
* Fix: error building my application #1414
* fix for vite
* Update vite.config.ts
* Update root.tsx
* fix the root.tsx and the debugtab
* lm studio fix and fix for the api key
* Update api.enhancer for prompt enhancement
* bugfixes
* Revert api.enhancer.ts back to original code
* Update api.enhancer.ts
* Update api.git-proxy.$.ts
* Update api.git-proxy.$.ts
* Update api.enhancer.ts
Added the new gemini-2.0-flash-thinking-exp-01-21 model to the GoogleProvider's static model configuration. This model supports a significantly increased maxTokenAllowed limit of 65,536 tokens, enabling it to handle larger context windows compared to existing Gemini models (previously capped at 8k tokens). The model is labeled as "Gemini 2.0 Flash-thinking-exp-01-21" for clear identification in the UI/dropdowns.
This PR introduces a new model, deepseek-r1-distill-llama-70b, to the staticModels array and ensures compatibility with the Groq API. The changes include:
Adding the deepseek-r1-distill-llama-70b model to the staticModels array with its relevant metadata.
Updating the Groq API call to use the new model for chat completions.
These changes enable the application to support the deepseek-r1-distill-llama-70b model, expanding the range of available models for users.
* feat: Integrate AWS Bedrock with Claude 3.5 Sonnet, Claude 3 Sonnet, and Claude 3.5 Haiku
* update Dockerfile for AWS Bedrock configuration
* feat: add new Bedrock model 'Mistral' and update Haiku to version 3
* feat: add new bedrock model Nova Lite and Nova Pro
* Update README documentation to reflect the latest changes
* Add the icon for aws bedrock
* add support for serialized AWS Bedrock configuration in api key
* fix: updated logger and model caching
* usage token stream issue fix
* minor changes
* updated starter template change to fix the app title
* starter template bigfix
* fixed hydretion errors and raw logs
* removed raw log
* made auto select template false by default
* more cleaner logs and updated logic to call dynamicModels only if not found in static models
* updated starter template instructions
* browser console log improved for firefox
* provider icons fix icons
