Repository navigation
v2.0.1
π DashScope SDK v2.0.1
This release introduces the DashScope 2.0 SDK stack alongside the existing 1.x API, bringing modern audio capabilities, improved CLI tooling, and enhanced documentation.
β¨ What's New
π Features
DashScope 2.0 SDK Stack
- Complete 2.0 SDK architecture landed alongside the legacy 1.x API
- New modular design with improved type safety and better developer experience
- Full compatibility maintained with existing 1.x code
- See the README for migration guidance
Audio API Preview
- 2.0 Audio API marked as preview with current speech models and protocols
- Text-to-Speech (TTS) with multiple voice options
- Speech-to-Text (ASR) with real-time streaming support
- Aligned with latest Alibaba Cloud speech service protocols
- Ready for testing and early adoption
π Improvements
CLI & Dependency Management
- Unified dependency installation prompts across all modules
- Simplified to single command:
pip install -U dashscope - Removed confusing
tokenizerandrlextra dependency groups - Cleaner
setup.pyconfiguration - Removed RL test options from
requirements-test.txt
- Simplified to single command:
Documentation
-
Comprehensive README updates for version 2.0
- Added feature comparison table between 1.x and 2.0
- Updated installation and usage examples
- New sections for audio, multimodal, and agent capabilities
-
Error handling documentation enhanced
- Marked error handling as work-in-progress in comparison table
- Added warning that 2.0 typed error system is under active development
- Provided forward-compatible exception handling recommendations
- Noted that new exception classes may be added and error code mappings adjusted
Examples & Samples
-
Reorganized examples directory with clear structure:
01_atomic_capabilities/- Core API demonstrations02_advanced_engineering/- Production patterns03_agent_recipes/- Agent development examples04_real_world_scenarios/- End-to-end use caseslegacy_1x/- 1.x API examples for reference
-
AgentStudio examples moved to 2.0 tier with agentic RL quickref
-
Legacy samples cleaned up and superseded
π Bug Fixes
- Fixed
agentstudiomemory update to send documented precondition object - Silenced
no-name-in-modulepylint warnings on Assistants imports - Improved code style compliance across examples
π¦ Package Information
- Version: 2.0.1
- Python: 3.9+
- License: Apache 2.0
- Platforms: POSIX, macOS, Windows
Installation
pip install -U dashscopeOptional Features
# Interactive AI assistant
pip install -U dashscope[acli]
# With Anthropic/OpenAI provider support
pip install -U dashscope[acli-anthropic]
pip install -U dashscope[acli-openai]
# Voice capabilities
pip install -U dashscope[acli-voice]
# Camera support
pip install -U dashscope[acli-camera]
# All optional features
pip install -U dashscope[acli-all]π Full Changelog
Commits: 14 commits since v1.27.8
feat(v2): land the dashscope 2.0 SDK stack alongside the 1.x API
feat(v2): mark the 2.0 audio API as preview
feat(v2.0.1): align audio with current speech models and protocols
refactor(cli): unify dependency installation prompts
docs(readme): update documentation to reflect new features and usage for version 2.0
docs(readme): update error handling documentation
docs(examples): move agentstudio to the 2.0 tiers, add agentic RL quickref
docs(examples): add the examples tree (legacy 1.x samples, agent recipes, real-world scenarios)
docs(samples): drop the last legacy sample
fix(agentstudio): send the documented precondition object on memory update
style(examples): silence no-name-in-module on Assistants imports
π Links
- PyPI: https://pypi.org/project/dashscope/2.0.1/
- Documentation: See README.md
- Issues: https://github.com/dashscope/dashscope-sdk-python/issues
π₯ Contributors
Thanks to all contributors who made this release possible: