Practical AI engineering
Practical articles, videos, SDK updates, architecture guidance, and tools for building AI applications with .NET.
Tracked .NET AI projects
Stable SDK and framework releases from selected .NET AI projects.
.NET: fix: race concdition in workflow formula state by @baywet in https://github.com/microsoft/agent-framework/pull/8252 .NET: update remaining dependencies by @baywet in https://github.com/microsoft/agent-framework/pull/8307 .NET: fix: inspect for delemiter value before adding…
api: add group with displayname to rate limits, deprecate grouptype (05fd63c) tools: add CompactBeforeNextTurn() to the tool runner (#239) (61cbe8b) client: don't replay a stream upload on the 401 token-refresh retry (#240) (2617360) client: join multiple anthropic-beta values…
See the official release notes for details.
api: add auto mode tool permissions for Managed Agents (0271c76) api: add compaction parameter and signed compaction blocks (beta) (8de3a4e) api: add enum types for workspace data-residency geo fields (59cd52e) api: add thinkingmismatchallowed entries to inputtransformations…
This month's release focuses on AI package reliability: closing gaps in evaluation scoring, hardening OpenAI image-option handling, and removing the deprecated OpenAI Assistants API support. OpenAI Assistants experimental APIs removed (was experimental under OPENAI001) #7724…
.NET: Replace deprecated AWSSDK.Extensions.Bedrock.MEAI with AWS.Bedrock.MEAI by @BenGearset in https://github.com/microsoft/agent-framework/pull/7983 .NET: Remove Azure.AI.OpenAI dependency by @rogerbarreto in https://github.com/microsoft/agent-framework/pull/7986 .NET: Track…
api: add auto mode tool permissions for Managed Agents (3b432a1) api: add contenttoolarge webfetch tool error code (e6b2b4f) api: add the user-profiles-2026-09-04 beta value and externaluserdetails to user profiles (04d3216) api: support mounting public GitHub repositories…
ONNX Runtime 1.30.0 expands generative AI inference, improves CPU and GPU performance, adds Go bindings, and strengthens runtime reliability. These notes cover changes since ONNX Runtime 1.29.1. - Expanded CUDA inference support with variable-length causal convolution for…
This is a patch release on top of v1.29.0, containing GroupQueryAttention capability and KV-cache layout improvements, plugin Execution Provider performance tooling updates, and targeted graph and optimizer fixes. - Added bidirectional GroupQueryAttention support on CPU and CUDA…
api: add named types for organization compliance settings state (34af6e4) api: add workspaceid request option for multi-workspace credentials (db8a192) api: raise Managed Agents vault refresh token limit to 8192 chars (62290cd) client: percent-encode path parameters and reject…
d8ec44919265b3641b7898a02038cbfe14590ee5 .NET: Bump package version to 1.80.1 (#14370) 3438d882147b6e3e19a161a5f0dcb64e23e181db .NET: update SDK to 10.0.303 (#14353) 437336175b8ef841a8ccace6c6cf909f34f0cad6 .NET: Remove retired OpenAI Assistants integration tests (#14354)…
This is a patch release on top of v1.28.1, containing a targeted fix for Compile API model serialization. - Fixed Compile API callback serialization to prevent duplicate graph nodes, inputs, outputs, and value information in emitted optimized models, including models with…
api: add externaluseronboardedat to user profiles (41cc8b1) api: add support for claude-fable-5-1 and claude-mythos-5-1 (80e4311) api: add support for organization compliance settings in the Admin API (7b89777) api: support pagination on organization rate limit list endpoints…
.NET: Bump AWSSDK.Extensions.Bedrock.MEAI from 4.0.6.10 to 4.0.101.8 by @dependabot[bot] in https://github.com/microsoft/agent-framework/pull/7829 .NET: Stabilize Foundry recovery tests by @rogerbarreto in https://github.com/microsoft/agent-framework/pull/7817 .NET: fix: preserve…
api: beta files/skills namespaces use GA shapes; drop dated beta header pins (6e61d8f) client: page backward via beforeid and stop on hasmore=false (cc8ef57) webhooks: require headers to be passed to unwrap() (ccd76b9) api: clarify pagination on the organization rate-limit list…
api: add updates thinking display mode (beta) (ae0cead) api: add missing anthropic-beta values (957f693) api: add support for Organization API endpoints (b255cba) tools: keep the tool runner going on pauseturn (#154) (44412b1) api: clarify session thread agent and…
streaming: keep accumulated text blocks wire-faithful when no citations stream (#156) (f9e1512) docs: clarify skill version latest support and memory version retention (131b16c)
Updated to llama.cpp 815a2a5915f22ce6a760c676389c5dfe8535c08f (b10221) by @martindevans in https://github.com/SciSharp/LLamaSharp/pull/1414 Implement Sequence ID pooling in BatchedExecutor to prevent native SeqMax overflow and crashes by @zsogitbe in…
.NET: Add session-persisted chat client routing by @westey-m in https://github.com/microsoft/agent-framework/pull/7641 .NET: Fix snakecase argument names in Harness file tool descriptions by @westey-m with @Copilot in https://github.com/microsoft/agent-framework/pull/7731 .NET:…
api: managed agents web search config and self hosted sandbox memory (c293479)
api: add outputbehavior to dream creation (create a new memory store or update the input store in place) (4b31214) api: additions to files and memory stores (9c9b7a6) api: Files and Skills APIs are now GA; add computer use and browser use toolsets (42206f3) api: updates to skill,…
This is a patch release on top of v1.28.0, containing support for device-free WebGPU compilation, improved compatibility with sandboxed Windows processes, and targeted graph-validation fixes. - Added support for device-free compile-only sessions, enabling offline graph…
5e1f1fb87d9a38ed44683228dda2434a9b2f0ef9 .NET: Bump package version to 1.80.0 (#14301) 1b7b020cd0a4e57d8d5d1cf2d515d352520ff424 Update OpenAPI HTTP client defaults (#14293) c1afc8639944e0e49c42aa796308ba2d0440f0d3 .NET: Bump Testcontainers packages (#14294)…
.NET: [Experimental] Extend A2A task store with isolation key scoping by @SergeyMenshykh in https://github.com/microsoft/agent-framework/pull/7504 .NET: Add CodeQL suppression for DevUI proxy validation by @SergeyMenshykh in https://github.com/microsoft/agent-framework/pull/7505…
This release adds hybrid stateful/stateless HTTP serving so clients using the 2025-11-25 and 2026-07-28 protocol revisions can share an endpoint, and fixes a malformed header-decoding edge case. Add HttpServerSessionMode for hybrid stateful/stateless HTTP serving #1796 by…
- onnxruntime-web has announced the deprecation of WebGL and JSEP. The native WebGPU EP is the recommended path going forward. See the deprecation and migration plans for details (#29716, #31683). - POSIX telemetry is now available on Linux, macOS, Android, and iOS when ONNX…
Version 10.9.0 is headlined by changes in these areas: AI: New experimental routing APIs center on the abstract RoutingChatClient base class, with SemanticRoutingChatClient as a concrete semantic-routing implementation. Separately, the abstract FailoverChatClient specialization…
See the official release notes for details.
api: add mid-conversation-tool-changes-2026-07-01 beta (1fcee67) api: add support for session budgets, advisor tool, pinned inference location and skills auto-loading from GitHub (1dc6838) client: preserve cache token counts when messagedelta omits them (#215) (75fbb01) api:…
fb137812c741b8d673ef331b2b517d028056b4bb .NET: Bump package version to 1.79.0 (#14275) 383d102346b7b29c929e0257ef672a48898b5f66 Bump .NET SDK from 10.0.301 to 10.0.302 (#14235) 362cc2bad7daec24d82562030358af755d581179 Bump form-data from 4.0.5 to 4.0.6 in…
This release adds an opt-in subscriptions/listen handler for custom server-side notification streams, improves HTTP transport fallback reliability, and expands guidance and samples for telemetry and binary resources. Keep AutoDetect channel open after provisional SSE failure…
.NET and Python: Extract Durable Task and Azure Functions integrations by @cgillum in https://github.com/microsoft/agent-framework/pull/7465 .NET: Fix Handoff orchestration sample not responding to user input by @peibekwe in https://github.com/microsoft/agent-framework/pull/7442…
.NET: fix InMemoryChatHistoryProvider persisting when service stores history by @westey-m in https://github.com/microsoft/agent-framework/pull/7284 .NET: Add TodoProvider and AgentModeProvider samples by @westey-m in https://github.com/microsoft/agent-framework/pull/7262 .NET:…
This servicing update refreshes the .NET AI project templates ahead of the July 30, 2026 retirement of GitHub Models — removing the GitHub Models provider option and updating template dependencies. As a result, both the AI Chat Web (aichatweb) and AI Agent Web API…
Version 2.0.0 brings the C# SDK into stable alignment with the MCP 2026-07-28 specification. This major release introduces discovery-first negotiation, multi-round-trip requests, stateless-by-default HTTP, caching hints, standardized headers, stronger OAuth and token-cache…
| Package | Version | |---|---| | Microsoft.Extensions.AI | 10.8.3 | | Microsoft.Extensions.AI.Abstractions | 10.8.3 | | Microsoft.Extensions.AI.OpenAI | 10.8.3 | - Updated serialization behavior for experimental ToolApprovalRequestContent.RequiresConfirmation so it no longer…
- Upgraded to ONNX 1.22.0 and protobuf 6.33.5 (#28754, #29606, #28967). Graph optimizer opset version checks were updated accordingly (#28966). - cuDNN and cuFFT are now optional at runtime for the CUDA EP, and nvrtc is no longer linked, which significantly reduces the required…
api: add claude-opus-5 model (bbee155) api: add tool addition/removal blocks and toolchange events (bbee155) api: expand client-side fallback credit token types and add server-side fallbacks default option (bbee155) client: streaming parallel tool-use emit logic (issue #197)…
Preserve message identity in MEAI chat streams (#359) by @awaescher in https://github.com/awaescher/OllamaSharp/pull/393 Full Changelog: https://github.com/awaescher/OllamaSharp/compare/5.4.29...5.4.30
Map embedding context size from MEAI options (#377) by @awaescher in https://github.com/awaescher/OllamaSharp/pull/392 Full Changelog: https://github.com/awaescher/OllamaSharp/compare/5.4.28...5.4.29
Fix model selection when unloading a model (#386) by @awaescher in https://github.com/awaescher/OllamaSharp/pull/391 Full Changelog: https://github.com/awaescher/OllamaSharp/compare/5.4.27...5.4.28
api: add new stop reason 'modelcontextwindowexceeded' (c7ee988)
This servicing release updates Microsoft.Extensions.VectorData.ConformanceTests to 10.8.2 and includes targeted test framework migration fixes. | Package | Version | Note | |---------|---------|---------| | Microsoft.Extensions.VectorData.Abstractions | 10.8.2 | Published August…
Python: Add MCPStreamableHTTPTool security guidance for custom http client by @TaoChenOSU in https://github.com/microsoft/agent-framework/pull/7245 Harden workflow credential selection by @moonbox3 in https://github.com/microsoft/agent-framework/pull/7249 Python: preserve Gemini…
api: add support for Managed Agents model effort, initial session events, and threads delta streaming (fdca188) helpers: restructure client-side refusal-fallback helpers (#106) (5fdc922)
api: add support for new refusal category (d9b672c) client: docs updates (19a59f0) docs: small updates (580c2cb) docs: small updates (86985e7)
.NET: Fix CosmosChatHistoryProvider: omit ttl when MessageTtlSeconds is nul… by @TheovanKraay in https://github.com/microsoft/agent-framework/pull/7030 .NET: Fix CompactionMessageIndex.IsSummaryMessage by @pwoosam in https://github.com/microsoft/agent-framework/pull/7042 .NET:…
This servicing release updates the Microsoft.Extensions.AI, Microsoft.Extensions.AI.Abstractions, and Microsoft.Extensions.AI.OpenAI packages to 10.8.1 with two targeted fixes: correct tool-call/tool-result ordering when resuming approval-gated functions with service-managed chat…
api: add support for dreaming (1228406) api: add support for MCP Tunnels (adea206) helpers: add session delta accumulator for agent.message previews (#78) (4b1f9d6) docs: updates to descriptions and examples (4696922)
Update package dependencies and align Microsoft.Extensions.AI by @awaescher in https://github.com/awaescher/OllamaSharp/pull/390 Full Changelog: https://github.com/awaescher/OllamaSharp/compare/5.4.26...5.4.27
Check .Message for null #387 by @awaescher in https://github.com/awaescher/OllamaSharp/pull/389 Full Changelog: https://github.com/awaescher/OllamaSharp/compare/5.4.25...5.4.26
This release adds new experimental APIs to Microsoft.Extensions.AI.Abstractions and updates the OpenAI dependency to 2.12.0, alongside documentation, test, and repository maintenance. New experimental API: AIFunctionNameAttribute and AIParameterNameAttribute #7610 by @jozkee…
This is a patch release on top of v1.27.0, containing targeted bug fixes, a CUDA QMoE decode-path optimization, and CI/build infrastructure fixes. - [MLAS] Fixed an igemm regression in the KleidiAI path (#28571) - Fixed a QMoE CPU livelock by eliminating nested intra-op…
This release backports a memory-leak fix for HTTP/SSE-based MCP servers. StreamableHttpServerTransport now releases its Server-Sent Events response stream reference as soon as a GET request ends, instead of holding it until the session is disposed via explicit DELETE or idle…
35ba23e1b3092271c778ca057afe1a796e16e70e .Net: Update package version to 1.78.0 (#14142) e6c9673684ca03621885083faff1644e1f42695e .Net: Disable automatic HTTP redirects in HttpPlugin and WebFileDownloadPlugin default clients (#14132) f25753be0a126138d2eea39b0ca2985252dde25a Bump…
.NET: Harden dotnet-format workflow shell handling by @SergeyMenshykh in https://github.com/microsoft/agent-framework/pull/6796 .NET: Add AgentSkillsSourceContext to AgentSkillsSource.GetSkillsAsync by @SergeyMenshykh in https://github.com/microsoft/agent-framework/pull/6797…
docs: small updates to field descriptions (8cc7e4b)
api: add agent-memory-2026-07-22 beta header (6e712be)
.NET: Fix typo in XML doc comment for workflow outputs param by @marcominerva in https://github.com/microsoft/agent-framework/pull/6326 .NET: Add BackgroundTaskCompletionLoopEvaluator for harness background agents by @westey-m in…
api: remove some nonfunctional types from the SDKs (7a70377)
Search and filter the complete 180-day archive of AI articles and videos.
Showing 24 of 117 stories
▶
45:18
Can AI actually handle on-call incident response, or is AIOps mostly hype? Alex Palcuie from Anthropic's AI Reliability team ...
▶
41:27
What I've learnt over a lot of customer and Microsoft's own AI ambition. ▭▭▭▭▭▭ C H A P T E R S ⏰ ▭▭▭▭▭▭ 00:00 ...
▶
8:26
Learn more about Valuemaxxing here → https://ibm.biz/~XUdepadwq For years, teams measured AI success through adoption ...
▶
0:52
Most developers build #RAG for semantic similarity. But similar ≠ relevant! Your coding agent can retrieve documents that match ...
▶
41:12
In this episode, Sahil Agarwal talks about the critical challenges of identity, authorisation, and security in the age of AI agents.
▶
Learn more about Agentic AI here → https://ibm.biz/~u8DMP3DGF Super agents can be powerful and dangerous. Grant Miller ...
▶
What happens when AI takes over incident response? Senior engineers build expertise through scar tissue from real production ...
Amazon SageMaker AI shipped 13 inference launches in year-to-date across two deployment paths: fully managed endpoints and Amazon SageMaker HyperPod Inference. This post reviews …
▶
Join Fiona and Robbie as they discus AI tokenomics with Jay Busch. Then watch Laura Salinas, Developer Advocate for Kiro, ...
▶
Your AI application needs useful context and a database that handles everyday traffic. See how Azure Cosmos DB stores ...
▶
Same question, two kinds of answer. An LLM writes a paragraph and then a boolean, and you get one bit you cannot threshold.
Kimi K3 from Moonshot AI is now available on Amazon Bedrock, giving you a powerful new open-weight option for coding and knowledge work. It offers native vision, a 1-million-token …
▶
IBM Research is helping shape the future of AI agents with open-source datasets, reinforcement learning environments, and ...
▶
Amanda Silver shares Microsoft's open source journey, from early ASP.NET support for jQuery to open-sourcing TypeScript, ...
Migrate a multi-model healthcare AI agent from self-managed Amazon ECS with AWS Fargate to Amazon Bedrock AgentCore runtime, preserving triple-model orchestration and …
Today we are announcing the new AgentCore runtime, a capability of Amazon Bedrock AgentCore built for the speed, flexibility, and cost efficiency that production agents demand. It …
Deploy production-ready Hugging Face models on Amazon SageMaker AI using six open-source agent skills. Point a coding agent at a model and get back a real-time endpoint with the …
Amazon SageMaker HyperPod Inference Gateway is a Kubernetes-native, GPU-aware routing add-on for Amazon EKS. It uses real-time GPU signals to send each inference request to the …
OpenAI introduces the Australian Youth Safety Blueprint, a six-pillar roadmap for safer AI experiences that protect and empower young people.
▶
What exactly is AI model distillation, and why does it become controversial when another company does it to your model? Ulysses ...
▶
Visit Mixture of Experts podcast page to get more AI content → https://ibm.biz/~pXHrXz07k On episode 125 of Mixture of Experts, ...
▶
Join the GameDay crew as we take on a real estate adventure for Unicorn.Rentals, building a team of AI agents live using ...
▶
Dario Amodei has called for "pacing the frontier" as increasingly capable AI systems are developed. But how practical is slowing ...
▶
Why should Australian businesses care about another Chinese AI model? Ulysses looks at DeepSeek V4.1 Flash and why an ...
No stories match these filters.
Official and practical resources
Start with the platform fundamentals, then choose the area that matches what you want to build.
Learn the .NET AI ecosystem, abstractions, SDKs, and application patterns.
Microsoft .NET AI documentation →Explore plugins, orchestration, memory, and agent-based application design.
Semantic Kernel documentation →Build Model Context Protocol clients and servers using the official C# SDK.
MCP C# SDK →Connect .NET applications to local models through Ollama without a hosted API.
OllamaSharp →