microsoft/semantic-kernel python-1.21.0 on GitHub

Release Notes

Add support for Azure AI Agent Service (#10414)
- Refer to the getting_started_with_agents/azure_ai_agent README for guidance on using Azure AI Agents in Semantic Kernel.
- Run your first sample to integrate an Azure AI Agent with a Semantic Kernel plugin.
Add Bedrock Agent (#10307)
- See the bedrock_agent README for setup instructions and usage details.
Add vector search to Postgres connector (#10213)
Document generator agent framework demo (#10184)
Implement MongoDB Atlas store (#10177)
Allow AI Inference connector to use Azure AI Services resource (#10427)
Allow factory callbacks in the process framework (#10451)
- See an example on how to configure a factory callback.
- An example of providing the factory to the ProcessBuilder can be found here.
Add Azure AI Agent AI Search Sample. Update memory sample with README. (#10497)
- Refer to the sample here.

Bump torch from 2.5.1 to 2.6.0 (#10380)
Update pymongo requirement from <4.11,>=4.8.0 to >=4.8.0,<4.12 (#10379)
Bump google-cloud-aiplatform from 1.60 to 1.79.0 (#10377)
Set openai package to newer version and remove default value of parallel tool calls (#10502)
Experimental Python 3.13 support (#10276)
Bump Python version to 1.21.0 for a release (#10500)

Introduce allowed content types in chat history channel receive. Add mixed chat image sample. (#10347)
Update README.md to add clarity regarding .env file (#10360)
Update document generator demo README (#10386)
Update agent tracing span name and attributes (#10398)
Improve streaming output formatting with code present and fix CSV file paths for learn site sample (#10498)
Improved content initializations, added ndarray support for binary content, and small fixes to defaults (#10469)
Remove old references of plugin import (#10462)
Fixes to Cosmos DB NoSQL query syntax generation (#10373)

Full Changelog: python-1.20.0...python-1.21.0