Windows 11 SCOOBE: Full-Screen Renewal Prompt in Insider Builds

Microsoft has begun shipping a significant Windows 11 Notepad update to Windows Insiders that embeds on‑device generative AI — adding Summarize, Write, and Rewrite actions that can run locally on qualifying Copilot+ PCs with no subscription required, while preserving a cloud fallback for subscribers who want higher‑capacity models.

Background​

Notepad has historically been the smallest, fastest plain‑text editor bundled with Windows, prized for its simplicity and near‑instantaneous launch. Over the last two years Microsoft has been incrementally modernizing its inbox apps, layering features like tabs, spell check, and basic AI helpers into previously minimal surfaces. The current Insider release formalizes a broader hybrid AI strategy for Windows: give qualifying devices a local inference path for routine generative tasks while continuing to offer cloud models for premium capacity and freshness.
The update — shipped as Notepad version 11.2508.28.0 — is rolling out to Windows Insiders in the Canary and Dev channels first. Microsoft’s official Windows Insider announcement details the new Notepad capabilities and confirms that local use on Copilot+ PCs does not require a Microsoft account or Microsoft 365 subscription; the cloud model remains available to subscribers who want it.

---

What’s included: Summarize, Write, Rewrite​

The three new actions — what they do​

• Summarize — Condenses selected text into short, medium, or long summaries. Useful for turning meeting notes, logs, or long drafts into concise action lists or abstracts.
• Write — Generates new text from a prompt or expands a selected fragment into paragraphs, lists, or templates. Ideal for quick email drafts, outlines, and rough content generation.
• Rewrite — Rephrases existing selections to change tone, formality, length, or clarity and returns alternative variants for quick polishing.

These actions are surfaced in Notepad via the Copilot menu, the right‑click context menu, and in current Insider flights sometimes via keyboard shortcuts and compact UI flows designed to keep users inside the editor.

User scenarios and practical value​

The intent is practical: give users a fast, in‑place writing assistant for short drafting, tone adjustments, and summarization so they don't have to switch apps or the browser for simple writing chores. For routine tasks — summarizing a chunk of meeting notes or polishing a paragraph before pasting into an email — local inference promises lower latency and better privacy posture than cloud-only alternatives.

---

The Copilot+ PC requirement and on‑device model basics​

What qualifies as a Copilot+ PC​

Notepad’s local AI path is gated to a device class Microsoft calls Copilot+ PCs — machines with an integrated Neural Processing Unit (NPU) capable of real‑time on‑device inference. Public guidance from Microsoft and corroborating reporting places practical NPU performance in the neighborhood of 40+ TOPS (tera‑operations per second) as the typical threshold for meaningful local generative workloads. Examples of silicon families that meet these thresholds include modern Snapdragon X‑series SoCs, certain Intel Core Ultra variants, and AMD Ryzen AI chips on qualifying SKUs.
This is a device‑level certification — OEMs must ship firmware/drivers and model provisioning flows that present the NPU to Windows as Copilot+ capable. Owners of older “AI‑branded” laptops should not assume compatibility without checking OEM documentation.

How local inference is managed​

When Notepad runs in local mode on a Copilot+ PC, the app uses a provisioned local model runtime and weight files that are downloaded and managed by Windows, the OEM, or Microsoft. The OS mediates model provisioning, storage, and updates; when both local and cloud are available, users with subscriptions can toggle between them. For users who are not signed in or lack a subscription, Notepad will fall back to the local model when available.

---

Rollout, language coverage, and availability​

The Insider blog entry confirms the feature set is initially available to Insiders on the Canary and Dev channels; Microsoft will stage broader availability thereafter. The initial on‑device model support is English only, with cloud models remaining the path for multilingual support and fresher web knowledge. Not every Insider on qualifying hardware will see the feature immediately — Microsoft gates flights by telemetry and device signals.
Practical steps to try the feature right now:

• Join Windows Insider (Canary or Dev).
• Confirm your device is Copilot+ certified with the OEM or device spec.
• Update Notepad through the Microsoft Store or wait for the inbox flight.
• Open Notepad, select text and use the Copilot menu or right‑click to choose Summarize, Write, or Rewrite.

---

Paint and Snipping Tool updates shipped in the same wave​

Microsoft shipped complementary updates to other inbox apps in the same Insider flight. Key changes include:

• Paint (version 11.2508.361.0) — Adds Project files (save and reopen working canvases) and an Opacity slider for brush and pencil tools.
• Snipping Tool (version 11.2508.24.0) — Introduces a Quick markup toolbar in the capture overlay that includes pen, highlighter, shapes, emojis, and integration buttons for Share, Visual Search with Bing, and Ask Copilot. Note: using some of those quick actions may change classic clipboard/autosave behaviors.

These non‑Notepad changes underscore Microsoft’s broader push to make inbox apps more capable and AI‑aware in small, incremental steps.

---

Why this matters: privacy, latency, and the subscription question​

Privacy: local inference reduces surface area​

Running inference on‑device means selected text does not need to leave the machine for many operations — a meaningful privacy win for sensitive notes or offline work. For organizations and privacy‑conscious users, local processing reduces transit exposure and potential cloud logging. However, local models still require provisioning and updates that may involve downloads and telemetry; administrators should inspect model provisioning flows and MDM controls.

Latency and responsiveness​

On‑device models can greatly reduce round‑trip latency compared to a cloud call, delivering near‑instant responses for short tasks like rewriting a sentence or creating a three‑line summary. That can materially improve the user experience in small, high‑frequency interactions.

Subscription economics and product strategy​

Microsoft’s hybrid approach decouples a baseline of capability from an explicit subscription barrier for qualifying hardware. By letting Copilot+ PC owners run local models without a Microsoft 365 subscription, Microsoft makes a pragmatic bet: differentiate by hardware capability while preserving cloud revenue opportunities for heavier usage and cross‑device continuity. For users without Copilot+ hardware, cloud models — and their associated sign‑in/credit policies — remain the path to advanced functionality.

---

What Microsoft has not (yet) disclosed — important technical unknowns​

Microsoft’s Insider announcement and public notes confirm the feature behavior and device gating, but they do not include granular technical specifications for the local Notepad model. Missing or under‑specified items include:

• Exact model architecture, family, or parameter counts for the local model.
• Precise NPU/CPU resource usage, or disk footprint of model weights on device.
• Formal benchmark data showing parity, latency, or accuracy comparisons between the local model and cloud models.

Because those details directly affect enterprise provisioning, capacity planning, and governance, they should be treated as unverified until Microsoft publishes technical documentation or independent third‑party benchmarks appear. Enterprises should pilot and measure behavior on representative Copilot+ hardware before adopting at scale.

---

Risks, limitations, and governance considerations​

Hallucination and content accuracy​

Generative models remain prone to producing plausible but incorrect information. Notepad’s outputs should be treated as assistive drafts — helpful for ideation and editing, not authoritative sources for legal, regulatory, or safety‑critical content. Implement review workflows when AI‑assisted content will be distributed externally.

Data loss, clipboard behavior, and DLP changes​

Quick markup in Snipping Tool changes the behavior of clipboard and autosave in some flows (for example, using Visual Search or Ask Copilot from the capture overlay may prevent automatic clipboard copies). Enterprises should test how these flows interact with existing Data Loss Prevention (DLP) rules and proxies. Notepad’s local processing reduces network egress for selected actions but does not remove the risk of data leaving the device if users invoke cloud features or other Copilot integrations.

Model provisioning, updates, and attack surface​

Local models still require binaries and runtime updates which are downloaded to endpoints. Those provisioning mechanisms create an operational surface that must be governed: who can allow models, how updates are validated, and what telemetry is collected. Administrators will want MDM controls to manage whether devices may provision local models and how change logs are exposed.

Hardware stratification and equitable access​

Tying full, subscription‑free local capability to Copilot+ hardware splits the Windows ecosystem. The result is a differentiated user experience that rewards newer, higher‑cost devices. For users on older hardware, cloud gating remains a potential friction point and cost consideration. This hardware stratification is both strategic (encouraging OEM innovation) and controversial (creating tiers of capability within the OS experience).

---

Recommendations for IT decision‑makers and power users​

• Inventory devices: identify which endpoints are Copilot+ certified and therefore eligible for local inference.
• Pilot deliberately: create a pilot group to evaluate performance, NPU utilization, thermal impact, and output quality on representative devices. Measure latency and battery behavior under realistic workloads.
• Update DLP policies: test Quick markup and Copilot integrations with existing DLP rules, and adjust policies for new clipboard/autosave behaviors.
• Control provisioning: use MDM controls to restrict whether local models may be provisioned or auto‑updated on managed devices. Demand vendor transparency on model change logs where possible.
• Educate users: communicate that AI outputs are suggestive drafts and require review before sharing externally; train staff on how to toggle local vs cloud modes if both are available.

---

Hands‑on: How to use Notepad’s AI features (Insider preview)​

• Join Windows Insider on the Canary or Dev channel and ensure Notepad is updated.
• Confirm your device is Copilot+ certified (check OEM specs).
• Open Notepad, select text and right‑click to choose Summarize, or open the Copilot menu to select Write or Rewrite.
• If you have a subscription, compare the local model’s output with the cloud model to decide which offers the best mix of speed and quality for your task.

If you prefer the original minimal Notepad, Microsoft provides settings to disable AI features and restore a classic, no‑frills editing surface.

---

Critical analysis: balancing product value and product theater​

Microsoft’s move to embed generative AI into Notepad is both technically notable and symbolically charged. On the technical side, bringing on‑device generative capabilities to a ubiquitous inbox app demonstrates how far NPUs and model optimization have progressed; it also validates Microsoft’s hybrid local/cloud architecture as a practical deployment pattern for desktop AI.
On the strategic side, grafting these features onto an app as minimal as Notepad is a clear signaling play: show AI everywhere and make even the smallest apps feel modern. That benefits power users who appreciate immediate drafting and editing assistance, but it also risks criticisms of feature‑bloat and confusing product positioning for users who prefer a lightweight editor. The hardware gating — Copilot+ certification — creates a split experience that both rewards OEMs and raises questions about equitable access to basic AI capabilities.
From an enterprise governance perspective, the update reduces some privacy concerns by enabling local inference but simultaneously raises new operational and compliance requirements related to model provisioning, telemetry, and DLP integration. The lack of detailed public model specs and benchmarks means IT teams must validate behavior empirically rather than relying on vendor claims alone.

---

What to watch next​

• Microsoft publishing detailed technical documentation: model families, weight sizes, and performance benchmarks for the on‑device Notepad models. Until those are published, claims of parity between local and cloud models should be treated cautiously.
• Expansion of language support and broader insider-to‑public rollout cadence for the feature. The initial English‑only rollout suggests a staged approach to localization.
• OEM documentation clarifying Copilot+ certification criteria and how device updates will be handled long term.
• Independent third‑party benchmarks comparing local vs cloud model accuracy, latency, and resource consumption across multiple Copilot+ hardware implementations.

---

Conclusion​

The Notepad update (version 11.2508.28.0) is a clear example of Microsoft’s dual strategy: democratize baseline AI capabilities with local models on Copilot+ PCs while offering premium cloud options to subscribers. For users on qualifying hardware, the change removes one friction point — the need to sign in or subscribe for small, high‑frequency generative tasks — and brings privacy and latency benefits. For IT teams and privacy‑conscious users, the change sharpens the focus on model provisioning, DLP, and the operational governance needed to manage local inference safely.
The update is available to Windows Insiders today; broader availability will follow as Microsoft tunes performance, telemetry, and controls. Users and administrators should pilot the capability, demand technical transparency where it matters, and treat AI outputs as assistive drafts rather than definitive answers.

---

Source: The Tech Outlook Microsoft releases a new update to Windows 11 Notepad app for Windows Insider; Generate, re-write and summarize text using on-device AI models - The Tech Outlook

Microsoft has quietly turned one of Windows’ simplest tools into a practical entry point for everyday generative AI: Notepad on Windows 11 can now write, rewrite, and summarize text using on‑device AI on qualifying Copilot+ PCs — and Microsoft is offering that on‑device path for free without a subscription for eligible hardware.

Background / Overview​

Notepad has not been immune to the AI wave sweeping desktop software. What began as a modest preview of a single “Rewrite” action has evolved into a trio of generative tools integrated directly into the classic Notepad experience: Write (generate new text from a prompt), Rewrite (rephrase and retone existing text), and Summarize (condense long passages). Those capabilities are surfaced in familiar places — the Copilot menu inside Notepad, the right‑click context menu, and accessible keyboard shortcuts in current Insider builds — keeping the experience intentionally low friction.
Official Windows Insider documentation published by Microsoft on September 17, 2025 confirms the rollout of these Notepad features to Windows Insiders in the Canary and Dev channels as part of Notepad version 11.2508.28.0, and it explicitly states that on‑device use is available on Copilot+ PCs without a subscription; subscribers can still choose to use cloud models when they want additional capability.
This is significant for two reasons. First, the local model path removes a paywall for basic AI assistance on qualifying hardware. Second, moving inference to the device prioritizes responsiveness and offers a different privacy trade‑off compared with cloud‑only workflows. Both points are central to Microsoft’s stated strategy of delivering hybrid AI across Windows inbox apps.

---

What’s new in Notepad: the three AI tools​

Write — generate drafts in place​

• What it does: Generates new content from a user prompt, or expands a small fragment into longer text (paragraphs, lists, templates).
• Where it appears: Copilot menu, right‑click menu, and in some Insider flights a keyboard shortcut (examples in current test builds include Ctrl + Q to open the Write dialog).
• Typical use cases: Quick email drafts, meeting notes expanded into action items, bullet lists turned into outline paragraphs, or drafting social posts without leaving Notepad.

Rewrite — polish tone and length​

• What it does: Rephrases selected text to change tone (casual → formal), shorten or lengthen passages, and propose alternate variants for selection.
• UI and controls: Users can right‑click a selection and choose Rewrite, then pick from suggested variants and tweak tone/length options. Earlier preview dialogs returned multiple variants so users could choose the best fit.
• Typical use cases: Converting a rough note into a more polished sentence, tailoring marketing copy for different audiences, or shortening text for microformats.

Summarize — compress long passages​

• What it does: Condenses highlighted text into short, medium, or long summaries; useful for turning meeting transcripts or long articles into readable abstracts.
• How it’s accessed: Highlight text → right‑click → Summarize, or use the Copilot menu and keyboard shortcuts found in Insider builds (some reporting shows Ctrl + M mapped to Summarize in test flights).
• Typical use cases: Extracting action items from logs, making long research notes scannable, or producing TL;DRs for quick sharing.

These actions are deliberately narrow in scope: Notepad is not being repositioned as a full writing studio but rather as an extremely low‑friction place to perform quick generative tasks without context switching. That design decision is important: the value proposition is speed and convenience more than deep composition or research‑grade output.

---

The hybrid model: local vs cloud, subscriptions and gating​

Microsoft’s messaging emphasizes a hybrid model:

• Local inference: On qualifying Copilot+ PCs, Notepad will use a local model runtime and weights to run Summarize/Write/Rewrite on‑device. Microsoft’s Insider post clarifies that local usage on Copilot+ hardware does not require a Microsoft account or Microsoft 365 subscription. This path is intended to be free and offline-capable.
• Cloud fallback / premium capacity: Users with Microsoft 365 or Copilot Pro subscriptions (or those signed into a Microsoft account where cloud models are allowed) can switch seamlessly to cloud models when they want higher capacity, fresher knowledge, or broader language support. Earlier Notepad previews required sign‑in and consumed cloud AI credits; that remains the cloud path for richer workloads.

Put simply: if your PC is Copilot+ certified and you’re on the new local model path, basic Notepad AI is available without subscription. If you need multilingual support, complex reasoning, or a more capable model, the cloud path — which may still require sign‑in and subscription credits in some regions — is the fallback. Independent reporting and third‑party coverage corroborate Microsoft’s hybrid framing.

---

How local AI is delivered and what “Copilot+ PC” means​

Microsoft’s local inference path depends on device‑level capabilities and provisioning:

• Copilot+ PCs are a certified device class that expose a Neural Processing Unit (NPU) or otherwise capable accelerator to Windows for on‑device model inference.
• The local model runtime and weight files are provisioned to the device by Windows/Microsoft or through the OEM; the OS mediates model storage and updates.

A few important technical notes and caveats:

• Microsoft has not published granular, public specs that tie a specific NPU throughput (e.g., exact TOPS numbers), model parameter counts, or memory footprints to Notepad’s local runtime. Community reporting and forum analysis have suggested practical performance thresholds (e.g., “40+ TOPS” cited as a ballpark for meaningful local generative work), but those numbers are not official Microsoft specifications and should be treated as unverified guidance until Microsoft publishes formal hardware requirements or independent benchmarks emerge.
• Because the device and model runtime handle inference locally, performance and responsiveness will vary by OEM device, NPU capability, RAM, and storage I/O. Real‑world behavior may favor higher‑end Copilot+ SKUs.

---

Rollout and language support​

• Current rollout: The Notepad AI update is being distributed initially to Windows Insiders in the Canary and Dev channels. Wider availability — to Beta/Release Preview and general consumers — is expected later as Microsoft stages flights. Not every Insider or Copilot+ PC will see the feature immediately; Microsoft gates flights by telemetry and device signals.
• Language coverage at launch: Local Notepad AI (the on‑device path) initially supports English only; cloud models remain the route for broader language support. Microsoft has signaled plans to expand languages over time.

---

Practical steps: how to try Notepad’s AI now​

• Join the Windows Insider Program and opt into the Canary or Dev channel.
• Confirm you have a Copilot+ PC (check your OEM documentation or device spec pages).
• Update Notepad via the Microsoft Store or wait for the inbox flight; verify Notepad version 11.2508.28.0 or later in the app’s About page.
• Open Notepad, highlight text, right‑click to access Summarize/Rewrite, or place the cursor and invoke Write from the Copilot menu (or use keyboard shortcuts available in current Insider builds).
• If you do not see the features, confirm your Insider channel, device certification, and that the build is installed; Microsoft gates availability by telemetry and device compatibility.

---

Strengths: why this matters for everyday Windows users​

• Accessibility of AI: By putting basic generative tasks into Notepad, Microsoft lowers the barrier for users who want quick drafts or a fast rewrite without opening a browser or a paid app. This democratizes basic AI writing assistance in a very practical way.
• Offline capability and privacy posture: Local inference reduces the need to send text to cloud services for routine tasks, which can be an advantage for private or sensitive notes. On‑device processing also reduces latency compared with cloud round trips.
• Familiar UX: Surfacing the features in the right‑click menu and Copilot menu keeps the integration lightweight and discoverable for users who already rely on Notepad as a quick scratchpad.
• Hybrid flexibility: Users who prefer or require higher capacity models can still switch to cloud models where available, preserving choice between free, offline local capabilities and subscription‑backed cloud models.

---

Risks and limitations — what to watch for​

• Model limitations and hallucinations: Local models optimized for on‑device performance will be smaller and may not match cloud models’ factuality or reasoning depth. Users must verify any factual claims or critical outputs before relying on them. This is particularly important for legal, financial, or medical content.
• Hardware stratification: The Copilot+ gating creates a split in user experience across the Windows ecosystem. Users on older hardware or non‑Copilot+ devices will rely on cloud models (and therefore sign‑in/subscription rules) or will not see the local path, which can be frustrating and raise fairness concerns.
• Enterprise controls and data loss prevention (DLP): IT teams must evaluate how Notepad’s AI integrates with organizational DLP and compliance policies. On‑device processing reduces cloud exposure, but enterprises still need to validate provisioning flows, model updates, and whether model caches or logs create new data leakage vectors.
• Undocumented resource impact: Microsoft has not published exhaustive documentation on the local model’s RAM, disk, and NPU usage. Organizations and power users should pilot the feature to measure battery, thermals, and performance impacts on their Copilot+ fleets before broad enablement.
• Language and regional limitations: Local models are English‑only at launch; users needing multilingual support must use cloud models that may require sign‑in or subscriptions. Expect staged expansion by Microsoft.

---

Verification and cross‑checks (what we confirmed)​

Several independent sources and Microsoft’s own Insider announcements corroborate the critical claims about Notepad’s new AI features:

• Microsoft’s official Windows Insider blog details the Notepad update (version 11.2508.28.0) and the inclusion of Summarize, Write, and Rewrite with local model support on Copilot+ PCs; it specifies English‑only support for the local models at launch and confirms the initial Insider Canary/Dev rollout.
• Reporting from Windows Central and BleepingComputer independently confirm that the on‑device path is being offered to Copilot+ PC owners without subscription gating, and they note the staged Insider rollout and language restrictions.
• Coverage by editorial outlets such as The Verge and Windows‑focused news sites documents UI elements (context menu access, Copilot menu), available keyboard shortcuts in test builds, and Microsoft’s broader strategy of augmenting inbox apps with practical AI features.

Where claims are unverified or clouded by incomplete documentation — for example, precise NPU performance thresholds and detailed model architecture for the on‑device runtime — community analysis and forum discussion provide helpful context but should be treated as provisional until Microsoft publishes formal hardware requirements or independent benchmarks appear.

---

Recommendations for users and IT administrators​

• For individual users who own a Copilot+ PC: test Notepad’s Summarize/Write/Rewrite in a safe environment first. Use the features for drafts and polishing, but always fact‑check AI outputs before using them in professional or high‑stakes contexts. If you are privacy‑conscious, local inference provides a reasonable first step, but verify how your device manages model weights and caches.
• For users without Copilot+ hardware: expect the cloud path (which may require a Microsoft account or subscription) to remain the primary option in the near term. Consider whether a subscription’s cloud capabilities are necessary for your workflows or whether you can wait for broader local device support.
• For IT teams and enterprise admins:
• Map which endpoints in your fleet are Copilot+ certified and prioritize pilot deployments there.
• Update device management policies to consider model provisioning, storage, and update flows (who controls the runtime, where weights are stored).
• Run a controlled pilot to measure performance, battery, and any unintended telemetry or logging behaviors.
• Review DLP and compliance posture; ensure use of Notepad’s AI aligns with your regulatory and data handling rules.

---

The broader picture: Notepad as an AI on‑ramp​

Microsoft’s choice to add free on‑device AI to Notepad on Copilot+ PCs is a pragmatic move that reflects two parallel trends: first, the steady push to embed modest AI helpers into everyday productivity surfaces; and second, the attempt to differentiate device classes by enabling offline AI on hardware certified for local inference. By bringing generative assist to Notepad — a tool almost every Windows user has opened at some point — Microsoft makes AI a less intimidating, more ubiquitous capability.
At the same time, the strategy introduces meaningful questions about device fragmentation, enterprise governance, and the limits of smaller on‑device models. Those trade‑offs will shape how quickly organizations and mainstream consumers adopt the local path versus sticking with cloud-based, subscription‑backed models.

---

Conclusion​

Turning Notepad from a scratchpad into a situational AI assistant is a subtle but consequential move. For owners of Copilot+ PCs, Microsoft is delivering free, offline, on‑device generative tools — Write, Rewrite, and Summarize — that make the simplest text editor on Windows an unexpectedly practical AI on‑ramp. The approach balances accessibility and privacy with the option to scale up via cloud models for users who want more capability.
This update is not an existential rewrite of desktop productivity, but it is a clear sign of how AI will be woven into the OS experience: small, practical features in familiar apps that incrementally reshape workflows. As the Notepad rollout expands beyond Insiders and local models gain broader language support, the real questions will be about governance, performance transparency, and whether Microsoft — and hardware partners — publish clearer specifications for on‑device AI so IT teams and users can make informed decisions. Until then, cautious experimentation on qualifying Copilot+ hardware is the sensible path forward.

---

Source: Cyber Press Microsoft Adds Free AI Support to Windows 11 Notepad