Back to Gap Map Product / UX / Orchestration & agents

Semantic Kernel

Microsoft
open source / Overall score: 3.7

Microsoft's SDK for building AI agents and integrating models into C#, Python or Java code. It exposes application code as model-callable plugins, often through OpenAPI, with connectors to AI services and vector stores, telemetry and responsible-AI hooks. Its first-class multi-language support distinguishes it from the Python-first frameworks.

Microsoft names the Microsoft Agent Framework as Semantic Kernel's successor, combining ideas from Semantic Kernel and AutoGen.

Openness

5 high confidence
5.0
license
MIT(OSI)
source
public(C#/Python/Java)
core features withheld
no

MIT multi-language SDK; first-class Azure connectors but model-agnostic and self-hostable. The repo page carries a banner that Semantic Kernel is now Microsoft Agent Framework, which is a succession notice rather than a change of terms.

  • https://github.com/microsoft/semantic-kernel recorded 2026-08-13

    MIT license badge and README statement "Licensed under the MIT license"; full C#/Python/Java source published, 28.4k stars and 4.7k forks; banner announcing Microsoft Agent Framework as the successor; no enterprise directory or paid tier

  • https://raw.githubusercontent.com/microsoft/semantic-kernel/main/LICENSE recorded 2026-09-29

    The repository LICENSE file is the MIT License: "MIT License", "Copyright (c) Microsoft Corporation." and a grant "to use, copy, modify, merge, publish, distribute, sublicense, and/or sell" the software.

  • https://raw.githubusercontent.com/microsoft/semantic-kernel/main/README.md recorded 2026-09-29

    The README calls Semantic Kernel "a model-agnostic SDK" with Python, .NET and Java support, states "Licensed under the [MIT](LICENSE) license." and mentions no enterprise edition, license key or paid tier (searched: enterprise edition, license key, paid: none).

Adoption

3 medium confidence
3.0

Adoption is measured by PyPI downloads of the semantic-kernel package over the trailing month, covering the Python distribution only. Because Semantic Kernel's .NET and NuGet channel is larger and not counted here, this reading is a floor rather than the full picture.

Capability

4 medium confidence
4.0

Semantic Kernel's strength is enterprise multi-language orchestration, with uniquely strong .NET and Java coverage beside the more common Python-first frameworks. That feature breadth matches AutoGen's.

Verified 2026-09-29, except some scores, still awaiting re-confirmation