Tessera
Folgen

Massimo Bonanni

@massimobonanni.bsky.social

"Paranormal Trainer, with the head in the Cloud and all the REST in microservices!" (cit.)

2821 Posts114 Follower68 Folgt
Massimo Bonanni@massimobonanni.bsky.social · 9h ago

Enterprise managed settings in GitHub Copilot for JetBrains

Enterprise managed settings in GitHub Copilot for JetBrains

GitHub Copilot for JetBrains now supports enterprise managed settings for plugin governance, MCP server access, OpenTelemetry, and permission modes. Administrators can now apply consistent controls for everyone on your enterprise’s… The post Enterprise managed settings in GitHub Copilot for JetBrains appeared first on The GitHub Blog.

github.blog

Massimo Bonanni@massimobonanni.bsky.social · 11h ago

Credential revocation and deauthorization by token type

Credential revocation and deauthorization by token type

Building on our self-service credential revocation experiences for incident response, you can now take token-type and user-specific actions to deauthorize and revoke user credentials during a security incident. This gives… The post Credential revocation and deauthorization by token type appeared first on The GitHub Blog.

github.blog

Massimo Bonanni@massimobonanni.bsky.social · yesterday

Microsoft named a Leader in the 2026 Gartner® Magic Quadrant™ for Cloud-Native Application Platforms

Microsoft named a Leader in the 2026 Gartner® Magic Quadrant™ for Cloud-Native Application Platforms

Cloud-native platforms are becoming the foundation for AI transformation. Discover how Microsoft's Azure application platform helps organizations modernize, innovate, and operate AI-powered applications at scale. The post Microsoft named a Leader in the 2026 Gartner® Magic Quadrant™ for Cloud-Native Application Platforms appeared first on Microsoft Azure Blog.

azure.microsoft.com

Massimo Bonanni@massimobonanni.bsky.social · 5d ago

Copilot Notebooks now works with Markdown, plain-text, and rich-text files

Copilot Notebooks now works with Markdown, plain-text, and rich-text files

Add your READMEs, wikis, logs, and transcripts to your notebook Your most useful project knowledge does not always live in a polished document. It might be captured in a README, project wiki, system log, meeting transcript, research export, or a simple text file. Copilot Notebooks can now use Markdown (.md), rich text format (.rtf) and plain text (.txt) files as references—helping you bring more of the content your work runs on into one focused workspace. Add the files to a notebook, chat with Copilot, and generate grounded insights and useful artifacts from your project context. Bring structured knowledge and raw data together Markdown often contains the structured knowledge behind a project: READMEs, technical documentation, project wikis, architecture notes, release notes, and runbooks. Plain-text files often contain raw data: system logs, exported transcripts, ticket dumps, customer feedback, and working notes. Rich Text Format (.rtf) files preserve formatted notes and documents—such as reports, drafts, and exported content. With support for Markdown, rich text, and plain-text files, Copilot Notebooks helps you bring together documentation, notes, logs, and transcripts into a single workspace for deeper insight and outcomes. Scenarios to try Here are a few ways to get started in notebooks: Turn a transcript into action: Add an exported meeting transcript and ask, “What did we decide, who owns each action, and what remains unresolved?” Find patterns in feedback: Add a plain-text export of support tickets and ask, “What are the most common issues, and which themes appear to be emerging?” Investigate an incident: Add a README and system log, then ask, “Compare the expected behavior with what happened before the failure.” Get up to speed on a new project: Add a project wiki, meeting transcripts, release notes, and working notes, then ask, “What do I need to know to contribute effectively to this project?” How it works Add a .md, .txt, or .rtf file as a reference in Copilot Notebooks, then chat with Copilot about its content. There is no need to reformat a log, convert a transcript into a traditional document, or manually summarize a long README first. Copilot can reason over the files in the context of the notebook and provide answers grounded in the source material. Feedback We want to hear from you. Try bringing a README, project wiki, log, transcript, or another text-based source into a Copilot Notebook—and let us know which scenarios help you move from complex files to useful work. More to come We’ll continue expanding the range of content you can bring into Copilot Notebooks—including new and emerging input types—so you can work with more of the context that matters to you. This feature is rolling out to all Copilot Notebooks users starting this week.

techcommunity.microsoft.com

Massimo Bonanni@massimobonanni.bsky.social · 5d ago

Updates to GPT-chat-latest in Microsoft Foundry

Updates to GPT-chat-latest in Microsoft Foundry

GPT-chat-latest in Microsoft Foundry is built on GPT-5.6 Sol, bringing focused responses, improved factual reliability, and consistent behavior to conversational applications without requiring developers to select a newly named model endpoint. This update gives developers access to the latest chat model improvements for advanced, natural, multimodal, and context-aware experiences while preserving an existing integration path for testing updated model behavior. What is new? Receive concise, relevant answers. GPT-chat-latest is now designed to respond more directly, use tighter formatting, and avoid superfluous details. For quick questions, applications can return concise answers with the context users need. For more involved tasks, including multi-step planning, research, and writing, the model can provide fuller responses while keeping the main recommendation clear. These response improvements are part of the latest GPT-5.6 Sol update. Improved factual reliability. GPT-chat-latest is designed to make fewer mistakes when a response depends on dates, numbers, sources, rules, or assumptions. This can help developers create conversational experiences grounded in the context supplied to the application. Teams should validate the improvement with representative prompts and factuality criteria before deployment. Built on GPT-5.6 Sol. GPT-chat-latest combines consistent behavior across straightforward questions and deeper tasks with GPT-5.6 Sol’s frontier reasoning capabilities for complex, multi-step work. What this means for developers Applications using GPT-chat-latest can benefit from these model improvements without requiring developers to select GPT-5.6 Sol by name. Because the underlying model can change as the endpoint is updated, teams should evaluate application behavior against their own prompts, data, tools, safety requirements, and quality thresholds before moving changes into production. Use cases Use GPT-chat-latest when a conversational application needs both direct, well-structured answers and advanced reasoning. It is a strong fit for workloads where the model must interpret detailed context, follow multiple constraints, and adjust response depth to the task, while giving developers an ongoing path to evaluate the latest chat model behavior through a stable endpoint. Customer support and self-service: Help users troubleshoot issues, understand product information, and navigate multi-step processes with direct answers grounded in approved knowledge sources. Planning and knowledge work: Break down complex objectives, reconcile constraints, and create detailed plans, briefs, or recommendations. Multimodal conversational experiences: Build applications that combine text and image context to answer questions, analyze visual information, and support richer interactions. Use GPT-chat-latest Use GPT-5.6 Sol Multi-turn assistants and customer-facing chat experiences that need natural conversation and consistent responses. Harder problems that benefit from more deliberate reasoning across multiple steps. Interactive experiences where users need quick back-and-forth clarification and task completion. Tasks involving multiple constraints, such as policy interpretation, detailed requirements, or long-horizon plans. Retrieval-augmented generation applications where the model decides when to retrieve information and synthesizes grounded answers. Offline or low-tool scenarios where the primary value comes from deeper reasoning over the context provided. Pricing Model Input ($/1M tokens) Cached input ($/1M tokens) Output ($/1M tokens) GPT-chat-latest $5$0.50$30 GPT-chat-latest is billed through your Azure subscription based on the deployment option and token usage. Standard deployments use pay-as-you-go pricing for input and output tokens, while provisioned throughput offers reserved capacity for workloads that need more predictable performance and costs. Pricing can vary by agreement, region, currency, and deployment type. Review the Azure OpenAI pricing page or use the Azure pricing calculator for current rates that apply to your configuration. Get started GPT-chat-latest in Microsoft Foundry, gives developers access to more focused responses, improved factual reliability, more consistent behavior, and advanced reasoning all without searching for a new model. Start by testing representative prompts and validating the updated model behavior against your application requirements. Try GPT-chat-latest in the Microsoft Foundry. Review the Chat Completions documentation for current implementation guidance before updating an application.

techcommunity.microsoft.com

Massimo Bonanni@massimobonanni.bsky.social · 6d ago

Routing and Failover for Microsoft.Extensions.AI

Routing and Failover for Microsoft.Extensions.AI

Route requests across models and providers natively in Microsoft.Extensions.AI with RoutingChatClient, SemanticRoutingChatClient, and FailoverChatClient — new experimental primitives for content-based routing, failover, and custom routing policies. The post Routing and Failover for Microsoft.Extensions.AI appeared first on .NET Blog.

devblogs.microsoft.com

Massimo Bonanni@massimobonanni.bsky.social · 6d ago

What 50 open source projects taught us about security in the AI era

What 50 open source projects taught us about security in the AI era

See how the open source projects in Session 4 of the GitHub Secure Open Source Fund combined AI-assisted workflows, maintainer expertise, GitHub security tools, expert guidance, and funding to improve project security. The post What 50 open source projects taught us about security in the AI era appeared first on The GitHub Blog.

github.blog

Massimo Bonanni@massimobonanni.bsky.social · 6d ago

Azure Content Understanding announces Synchronous Operations

Azure Content Understanding announces Synchronous Operations

Workflow automation scenarios—including grounding AI agents, verifying identities, assisting customers with documents in call centers, and triggering document-based workflows—depend on immediate content processing, where every second matters. Now in public preview, synchronous Read, Layout, and Digital Parse operations in Azure Content Understanding, part of Foundry Tools, deliver real-time content extraction without temporary service-side storage. Note: The features, limits, pricing, API versions, and SDK details described in this post apply to a public preview and may change before general availability. Why are we introducing synchronous operations? Generative AI–powered agentic systems increasingly rely on real-time content extraction from images and documents. These agents often need information as soon as a document enters a workflow. For example, when a customer uploads a document during a call center chat, synchronous Read and Layout operations can immediately return structured content for the agent to use in its response, eliminating the delay of an asynchronous job. Enterprise customers using Azure Document Intelligence and Azure Content Understanding (part of Foundry Tools) have asked for an option that avoids temporary service-side storage when processing highly confidential documents. The new synchronous Read and Layout operations address this requirement by processing documents in memory without temporary storage. Each request can provide the document as binary data or through an accessible URL; the service extracts the text and structure and returns structured content directly in the response. This capability complements Azure Content Understanding’s standard asynchronous operation, which has been generally available since November 2025. What do synchronous operations provide? In this preview, the prebuilt Read (prebuilt-read), Layout (prebuilt-layout), and Digital Parse (prebuilt-digitalParse) analyzers support both asynchronous and synchronous operations. All other prebuilt analyzers continue to support asynchronous operations only. Read uses an OCR model to extract text and its location information from images and PDF documents. Layout apart from extracting text and its location information, takes steps further to extract document structure such as tables, sections, figures, formats, hyperlinks and now signatures from more than 30 file types. Synchronous operations support documents up to 10 MB and process up to five pages or 30,000 characters per request. You can optionally specify the page range to process. Synchronous operations support the same document file formats as asynchronous operations. For complete service limits, refer to Azure Content Understanding documentation. Figure: Azure Content Understanding synchronous operations enable low-latency content extraction from documents and images, returning structured results directly without temporary service-side storage. Synchronous requests use priority processing and resources designed to support lower-latency scenarios. Although they are priced at a premium relative to asynchronous operations, they remain competitively priced, as shown below. For complete pricing information, refer to Azure Content Understanding pricing. Capability Pricing meter Asynchronous operation Synchronous operation (preview) Read: Extract text from images and PDFs using OCR. Document: Basic $1.00 per 1,000 pages $1.50 per 1,000 pages Layout: Extract content from images and PDFs while identifying document structure, sections, formatting, figure types, hyperlinks, and signatures. Document: Standard $5.00 per 1,000 pages $7.50 per 1,000 pages Text documents: Extract content from supported text-based files. Document: Minimal $0.01 per 1,000 pages* $0.015 per 1,000 pages * Up to 3000 characters counted as one page (round up). How can I try and adopt Azure Content Understanding? You can try and adopt Azure Content Understanding’s synchronous operations using REST API calls, or the Azure Content Understanding SDK. Before you begin, make sure you have an active Azure subscription and a Microsoft Foundry resource. Synchronous operations can be invoked by two methods (REST: analyzeInline / analyzeBinaryInline): analyze_inline — input content as a URL analyze_binary_inline — input content as a binary file curl commands Sample curl command to extract content from an image binary using the synchronous Read operation: curl --request POST \ --url 'https://{your-resource-endpoint}/contentunderstanding/analyzers/prebuilt-read%3AanalyzeBinaryInline?api-version=2026-06-01-preview' \ --header 'Content-Type: application/octet-stream' \ --header 'Ocp-Apim-Subscription-Key:{your-subscription-key}' \ --data-binary '@D:\Demo\InsuranceCard.png' Sample curl command to extract content from a document URL using the synchronous Layout operation: curl --request POST \ --url 'https://{your-resource-endpoint}/contentunderstanding/analyzers/prebuilt-layout%3AanalyzeInline?api-version=2026-06-01-preview' \ --header 'Content-Type: application/json' \ --header 'Ocp-Apim-Subscription-Key:{your-subscription-key}' \ --data '{ "inputs": [ { "url": "https://github.com/Azure-Samples/azure-ai-content-understanding-python/raw/refs/heads/main/data/invoice.pdf" } ] }' Code samples using Azure Content Understanding SDK The Azure Content Understanding SDK has been updated to support synchronous Read and Layout operations. You can find links to the SDKs in the References section below. The following code sample uses the Azure Content Understanding SDK in Python programming language. Install the Azure Content Understanding SDK supporting preview features: python -m pip install --pre azure-ai-contentunderstanding Sample using the SDK to extract content from an image binary with the synchronous Read operation: import json from azure.ai.contentunderstanding import ContentUnderstandingClient from azure.core.credentials import AzureKeyCredential endpoint = "https://{your-resource-endpoint}" credential = AzureKeyCredential("{your-subscription-key}") client = ContentUnderstandingClient(endpoint=endpoint, credential=credential) with open(r"D:\Demo\InsuranceCard.png", "rb") as f: file_bytes = f.read() inline_response = client.analyze_binary_inline( analyzer_id="prebuilt-read", binary_input=file_bytes, ) with open("read_result.json", "w", encoding="utf-8") as f: json.dump(inline_response.result.as_dict(), f, indent=2) Sample using the SDK to extract content from a document URL with the synchronous Layout operation: import json from azure.ai.contentunderstanding import ContentUnderstandingClient from azure.ai.contentunderstanding.models import AnalysisInput from azure.core.credentials import AzureKeyCredential endpoint = "https://{your-resource-endpoint}" credential = AzureKeyCredential("{your-subscription-key}") client = ContentUnderstandingClient(endpoint=endpoint, credential=credential) inline_response = client.analyze_inline( analyzer_id="prebuilt-layout", inputs=[ AnalysisInput( url="https://github.com/Azure-Samples/azure-ai-content-understanding-python/raw/refs/heads/main/data/invoice.pdf" ) ], ) with open("layout_result.json", "w", encoding="utf-8") as f: json.dump(inline_response.result.as_dict(), f, indent=2) References User documentation Pricing Azure Content Understanding SDK for Python Azure Content Understanding SDK for .NET Azure Content Understanding SDK for JavaScript Azure Content Understanding SDK for Java Send your feedback to cu_contact@microsoft.com

techcommunity.microsoft.com

Massimo Bonanni@massimobonanni.bsky.social · 6d ago

Built for business: How Microsoft 365 Copilot keeps you in the flow of legal work

Built for business: How Microsoft 365 Copilot keeps you in the flow of legal work

The tools that run a business rarely live in one place. For a small business owner, that means stepping out of the document you're writing to search for an attorney, digging through a folder for the LLC paperwork you half-remember filing, or pausing a client deliverable to figure out if a contract clause is enforceable. Every one of those detours pulls you out of the work you were doing to go chase down something that should've been a quick answer. For a small business, where focus is one of the most valuable resources there is, constant switching adds up. Microsoft 365 Copilot is designed to keep your work moving, bringing your data, expertise, and workflows into the apps you already use like Word, Excel, PowerPoint, Outlook, and Teams - now powered by AI. And through built-in and custom agents, Copilot extends to include specialized tools and services that run your business, so you can use their capabilities all in one place. That ecosystem of partner agents is growing quickly, and LegalZoom is a great example of what it makes possible: chat with a LegalZoom Legal Assistant, get a recommended business formation plan, or connect with an experienced attorney - all without needing to leave Copilot.* Handle legal tasks with LegalZoom, directly in Microsoft 365 Copilot With the LegalZoom agent, resources are provided to help answer a legal question the moment it comes up, right inside Microsoft 365 Copilot. Instead of jumping out to another tool, the work continues where you already are. And when a task calls for something more, the agent provides a streamlined path back to LegalZoom, which remains the home for its services and the professionals behind them. That’s the shift that agents make possible in Microsoft 365 Copilot: the tools a business depends on coming to where work is already happening, rather than living in separate siloes. “Small business owners shouldn’t have to leave their work environment to get reliable legal guidance. With LegalZoom inside Microsoft 365 Copilot, we are helping entrepreneurs find answers, understand their options and connect with attorney support at the exact moment they need it.”* — Jeff Stibel, Chairman & CEO, LegalZoom *Microsoft 365 Copilot does not provide legal advice or legal services. Built on a foundation of trust Bringing more of your business into Microsoft 365 Copilot only works if it’s secure. Microsoft 365 Copilot provides enterprise-grade security, privacy, and compliance tools and it honors the access permissions your business has already established. Your data remains within your organization's Microsoft 365 security and permissions boundaries. Microsoft does not use your prompts, responses, files, emails, chats, or other Microsoft Graph data accessed by Microsoft 365 Copilot to train the foundation LLMs used by Microsoft 365 Copilot. And we don't sell your information to advertisers. Stay tuned for more in this series and learn more about Microsoft 365 Business with Copilot. The promise is simple: the tools and expertise a business relies on, brought into the flow of work rather than scattered across it. The LegalZoom agent is available now through the Microsoft Marketplace. To learn more, visit legalzoom.com. About LegalZoom LegalZoom is a leading online platform for legal services, transforming how individuals and small businesses navigate the legal system. By combining intuitive technology with access to experienced attorneys through their vast independent attorney network or their own law firm, they offer the tools and guidance people need to confidently manage everything from LLC formation and compliance to intellectual property protection and ongoing business management and legal support. For more information, please visit legalzoom.com.

techcommunity.microsoft.com

Massimo Bonanni@massimobonanni.bsky.social · last wk.

Azure Developer CLI extension framework is GA: build dev workflows for apps using Azure

Azure Developer CLI extension framework is GA: build dev workflows for apps using Azure

What parts of building and shipping on Azure still take too many steps? Maybe you want a command that starts a project with your organization’s approved architecture and infrastructure. Maybe you need to connect deployments to an internal service catalog, run security checks before provisioning, or guide developers through a process that currently lives across […] The post Azure Developer CLI extension framework is GA: build dev workflows for apps using Azure appeared first on Azure SDK Blog.

devblogs.microsoft.com

Massimo Bonanni@massimobonanni.bsky.social · last wk.

Automatically migrate branch protection rules to repository rulesets

Automatically migrate branch protection rules to repository rulesets

You can now convert your existing branch protection rules to repository rulesets directly from your repository settings. This helps you migrate to GitHub’s more flexible and scalable policy framework without… The post Automatically migrate branch protection rules to repository rulesets appeared first on The GitHub Blog.

github.blog