inference_provider_sdk 1.0.2
inference_provider_sdk: ^1.0.2 copied to clipboard
Flutter/Dart SDK for Inference Provider API - Build powerful AI agents with RAG, tool calling, and MCP integration
Changelog #
All notable changes to this project will be documented in this file.
The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.
1.0.2 - 2025-01-11 #
Changed #
- Made repository public to enable full pub.flutter-io.cn analysis
- Improved example package structure with dedicated
pubspec.yamland README - Enhanced documentation for better pub.flutter-io.cn scoring
Fixed #
- Fixed pub.flutter-io.cn incomplete analysis issue by making repository publicly accessible
- Improved example package structure to meet pub.flutter-io.cn best practices
1.0.1 - 2025-01-11 #
Fixed #
- Fixed static analysis issues to pass pub.flutter-io.cn scoring (0/50 → 50/50)
- Removed deprecated lint rules (
always_require_non_null_named_parameters,prefer_equal_for_default_values) - Converted exception constructors to use super parameters for modern Dart syntax
- Removed unnecessary library name declaration
- Added
flutter_lintspackage to dev dependencies - Excluded example directory from analysis to allow
printstatements in examples
Changed #
- Updated
analysis_options.yamlto comply with latest Dart best practices - Limited package topics to 5 (pub.flutter-io.cn requirement)
1.0.0 - 2025-01-11 #
Added #
- Initial release of the Inference Provider Flutter/Dart SDK
- Agent Management:
- Full support for Agent CRUD operations
- Agent inference with text and vision capabilities
- Conversation history management
- Tool Management:
- Create, read, update, delete tools
- Support for REST API tools and JavaScript functions
- Attach/detach tools to/from agents
- Priority-based tool ordering
- MCP Server Integration:
- Full MCP server management (CRUD)
- Attach/detach MCP servers to/from agents
- Priority-based server ordering
- Support for MCP server commands and arguments
- RAG / Vector Search:
- Upload files for vector search
- Search vectors with similarity thresholds
- Manage RAG files per session
- Configurable chunking (size and overlap)
- AI Provider and Model Management:
- Manage AI providers and models
- Support for multiple provider types
- Error Handling:
- Comprehensive exception hierarchy:
AuthenticationExceptionValidationExceptionNotFoundExceptionRateLimitExceptionServerExceptionNetworkException
- Full JSON serialization with
json_serializable - Type-safe models with
json_annotation - HTTP client with automatic error handling
- Variable substitution support
- Complete Dart documentation
- Example Flutter application
Dependencies #
- Dart SDK: >=3.0.0 <4.0.0
- http: ^1.1.0
- json_annotation: ^4.8.1
Features Overview #
- ✅ Full Dart type safety with JSON serialization
- ✅ Agent inference with conversation history
- ✅ RAG (Retrieval-Augmented Generation)
- ✅ Tool calling (REST API, JavaScript, MCP)
- ✅ MCP server integration
- ✅ Provider and model management
- ✅ Custom response formatting
- ✅ Comprehensive error handling
- ✅ Vision support for image-based inference