Recommended Free Tools
iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
“Local AI” tells you where a model performs inference; it does not prove that setup, surrounding code context, telemetry, logs, or a fallback request stay on the device. To know what leaves a particular app, you need its exact version and configuration plus evidence of the route it actually took. The documented examples below explain what to check—not a firsthand trace of any one app.
What “local” does—and does not—tell you
Local inference means the model computes on the device. It is only one stage in an AI feature’s data flow. An app may need a network connection to obtain a model, may refresh catalog information, may send context to an AI provider, or may offer a separately configured cloud fallback. Those paths are distinct; the existence of one does not establish that the others are active.
Microsoft’s hybrid-design guidance recommends checking whether a local model is ready and using a cloud endpoint only when the user and organization permit data to leave the device. If cloud use is not permitted, the app should explain the requirement and disable or hide the feature. Microsoft also recommends recording which route was used and readiness or fallback errors without recording prompts or sensitive content unless approved. Microsoft’s cloud-versus-local AI guidance describes that design approach.
Separate the stages of the data flow
Model download and catalog updates
Microsoft says Foundry Local runs inference on-device once a model has been downloaded and cached. Its initial model download requires internet access; optional model-catalog refreshes may also occur, while cached catalog information can support offline inference. Microsoft summarizes Foundry Local’s network activity this way: “The only network traffic is the initial model download and optional catalog metadata refreshes.” That statement concerns Foundry Local, not every feature described as local. Microsoft’s Windows AI FAQ also says Foundry Local can select a Qualcomm NPU, DirectX 12 GPU through WinML/DirectML, NVIDIA GPU through CUDA, or CPU fallback for execution.
#1 Best Overall
- EVOLUTION AMD RYZEN AI MAX+ 395 MINI PC - GMKtec EVO-X2 is the next evolution in AI mini PC Ryzen Strix Halo series. Thanks to AMD Simultaneous Multithreading (SMT) the core-count is effectively doubled, to 32 threads. Ryzen AI Max+ 395 has 64 MB of L3 cache and can boost up to 5.1 GHz, depending on the workload. The Ryzen AI Max+ 395 is currently rated as the "most powerful x86 APU" on the market for AI computing.
- AI NPU with XDNA 2 ARCHITECTURE - Powered by 16 “Zen 5” CPU cores, 50+ peak AI TOPS XDNA 2 NPU and a truly massive integrated GPU driven by 40 AMD RDNA 3.5 CUs, the Ryzen AI MAX+ 395 is a transformative upgrade and delivers a significant performance boost over the competition. The Ryzen AI Max+ 395 excels in consumer AI workloads like the llama.cpp-powered application: LM Studio. Shaping up to be the must-have app for client LLM workloads, LM Studio allows users to locally run the latest language model without any technical knowledge required and unleash their creativity and productivity.
- AMD RADEON 8090S iGPU GAMING PC - The AMD Radeon RX 8060S offers all 40 CUs with up to 2.9 GHz graphics clock and uses the new RDNA 3.5 architecture. The powerful iGPU is positioned between an RTX 4060 and 4070 laptop GPU and therefore enables gaming in FHD at maximum details in most demanding games. The 8060S can also utilize the full 128GB pool, which is perfect for running LLMs such as Deepseek 70B Q8, which runs comfortably on this machine.
- EIGHT CHANNEL LPDDR5X - LPDDR5X is a new ground breaking memory small form factor installed on-board. With blazing speeds up to to 8000MT/s, it runs 1.5x faster than the DDR5 SODIMMs; 90% better performance over DDR5 SODIMMs in video conferencing and photo editing; 30% better performance in productivity apps; 12% better performance in digital content workloads.
- QUAD SCREEN 8K DISPLAY SUPPORT - EVO-X2 AI Mini PC support 4-screen 4K/8K output via HDMI 2.1 (8K@60Hz), DisplayPort 1.4 (4K@60Hz), and dual USB 4 40Gbps Transfer speed (supporting PD3.0/DP1.4/DATA). Ideal for gaming, video editing, and multitasking, it provides expansive and crisp multi-display support.
Inference request and fallback
A local model may be unavailable because it is not installed, is not ready, or does not support the requested capability. An app might then refuse the request, ask for permission to download or change routes, or use a cloud endpoint if that is allowed by its configuration. These are possible design states, not proof that an unspecified app silently sends a request to the cloud. The key questions are what triggers fallback, whether the route is fail-closed or fail-open, and whether the user or administrator can block cloud use.
Prompt context
The data sent for a coding request can exceed the text a user typed. For Gemini Code Assist Standard and Enterprise, Google lists conversation history, snippets from open files, snippets from adjacent files, and cursor location among the Customer Data that may be processed. Google says the service is stateless and does not store prompts and responses in Google Cloud by default, though customers can configure Cloud Logging. Its documentation says processing typically occurs at the closest data center but does not guarantee regional processing. These details apply to the specified Gemini Code Assist editions; they do not establish another product’s context handling or processing location. Google’s security, privacy, and compliance documentation sets out that scope.
Rank #2
- Built for Local AI Development: AMD Ryzen AI Halo is designed for local AI development and inference, featuring 128GB unified memory and support for up to 200B parameter models to build and run intensive AI workloads locally.
- 128GB Unified Memory: Features 128GB LPDDR5x unified memory at 8000 MT/s with 256 GB/s memory bandwidth, providing a shared memory pool across the CPU, GPU, and NPU to support larger AI models.
- AMD Ryzen AI Max+ 395 Processor: Features 16 cores, 32 threads, and Zen 5 architecture, paired with AMD Radeon 8060S integrated graphics featuring 40 RDNA 3.5 compute units and an AMD XDNA 2 NPU with up to 50 TOPS.
- Linux AI Developer Platform: Purpose-built for Linux-based AI development with full AMD ROCm software support and preloaded tools, models, and workflows optimized for local AI development.
- Compact, Connected Design: Includes a 2TB M.2 SSD, 10GbE LAN, Wi-Fi 7, Bluetooth 5.4, USB-C connectivity, and HDMI 2.1b.
JetBrains documents that AI Assistant may send code fragments and context such as file types and frameworks to its LLM provider. Its 2026.2 documentation also describes opt-in detailed AI usage collection, which includes full communication and is disabled by default in the documentation reviewed, and a session request log named ai-assistant-requests.md that users can inspect. Check the installed version and license settings rather than assuming those defaults describe every installation. JetBrains’ AI Assistant data-handling documentation explains the product-specific details.
Free tools Windows power users keep installed
One-click scans. No signup required.
Telemetry and content logs
Telemetry can record events without recording the request itself. Google’s Gemini Code Assist documentation gives examples such as a request event without request contents, a response event, user reactions, accepted-suggestion character counts, and interface interactions. Google says engineers can access this telemetry to support product improvements. Google’s telemetry documentation describes those examples.
Rank #3
- EVOLUTION AMD RYZEN AI MAX+ 395 MINI PC - GMKtec EVO-X2 is the next evolution in AI mini PC Ryzen Strix Halo series. Thanks to AMD Simultaneous Multithreading (SMT) the core-count is effectively doubled, to 32 threads. Ryzen AI Max+ 395 has 64 MB of L3 cache and can boost up to 5.1 GHz, depending on the workload. The Ryzen AI Max+ 395 is currently rated as the "most powerful x86 APU" on the market for AI computing.
- AI NPU with XDNA 2 ARCHITECTURE - Powered by 16 “Zen 5” CPU cores, 50+ peak AI TOPS XDNA 2 NPU and a truly massive integrated GPU driven by 40 AMD RDNA 3.5 CUs, the Ryzen AI MAX+ 395 is a transformative upgrade and delivers a significant performance boost over the competition. The Ryzen AI Max+ 395 excels in consumer AI workloads like the llama.cpp-powered application: LM Studio. Shaping up to be the must-have app for client LLM workloads, LM Studio allows users to locally run the latest language model without any technical knowledge required and unleash their creativity and productivity.
- AMD RADEON 8090S iGPU GAMING PC - The AMD Radeon RX 8060S offers all 40 CUs with up to 2.9 GHz graphics clock and uses the new RDNA 3.5 architecture. The powerful iGPU is positioned between an RTX 4060 and 4070 laptop GPU and therefore enables gaming in FHD at maximum details in most demanding games. The 8060S can also utilize the full 64GB pool, which is perfect for running LLMs such as Deepseek 32B, which runs comfortably on this machine.
- EIGHT CHANNEL LPDDR5X - LPDDR5X is a new ground breaking memory small form factor installed on-board. With blazing speeds up to to 8000MT/s, it runs 1.5x faster than the DDR5 SODIMMs; 90% better performance over DDR5 SODIMMs in video conferencing and photo editing; 30% better performance in productivity apps; 4% better performance in digital content workloads.
- QUAD SCREEN 8K DISPLAY SUPPORT - EVO-X2 AI Mini PC support 4-screen 4K/8K output via HDMI 2.1 (8K@60Hz), DisplayPort 1.4 (4K@60Hz), and dual USB 4 40Gbps Transfer speed (supporting PD3.0/DP1.4/DATA). Ideal for gaming, video editing, and multitasking, it provides expansive and crisp multi-display support.
Prompt and response logging is a separate choice. Google documents optional Cloud Logging for prompts and responses, context, and metadata such as telemetry and accepted lines of code. When enabled, data is sent to Cloud Logging for organization administrators. Google says prompts and responses are not used to train its model whether or not logging is enabled. Do not treat that statement about training as a claim that no data is transmitted or logged. Google’s logging documentation explains the configurable logs.
Provider routes and proxies
Some products insert a proxy or intermediary into the request path. Codag’s documentation, for example, says its local proxy routes model traffic directly to the user’s provider, while eligible large tool outputs plus minimum task context may go to Codag for transient processing. It says source code, diffs, configuration, and unrecognized content pass through unchanged, and describes operational metrics as contentless. These are Codag’s own claims about its product, not independent validation or a general description of local AI tools. Codag’s privacy and data-flow documentation states its account.
Rank #4
- 【Leading AI Mini Workstation】MINISFORUM AI MS-S1 Max Workstation comes with AMD Ryzen AI Max+ 395 processor, which uses AMD's latest generation Zen 5 architecture. It has 16 Cores and 32 Threads, the boost clock is up to 5.1GHz. The overall processor performance is up to 126 TOPS, and the NPU performance reaches up to 50 TOPS. AMD Ryzen AI enables improved productivity, advanced collaboration, and improved efficiency.
- 【AMD Radeon 8060S Graphics 】The MS-S1 Max Mini PC equipped with AMD Radeon 8060S Graphics which built on the new generation of RDNA 3.5 architecture AMD graphics, it brings ultra-high frame rate experiences and advanced content creation features anywhere and delivers staggering performance. It can handle all your computing and multimedia tasks efficiently.
- 【Five 8K Video Output】This MS-S1 Max Workstation comes with five video outputs, 1x HDMI (8K@60Hz), 2x USB4(40Gbps,Alt DP2.0,PD out 15W) and 2x USB4 V2(80Gbps,Alt DP2.0,PD out 15W) Outputs, which support multiple monitors display at the same time and provide a larger and wider filed of view and improve your work efficiency. It is used in fields that require high-performance computing and graphics processing, including digital signage and securities trading, as well as work that uses CAD, such as engineering design, scientific calculations, animation production, and post-production for movies and television
- 【 Fast and Stable Wire & Wireless Speed】It comes with Two 10G Lan Ports for wired connection and and Wi-Fi 7 / BT5.4 for wireless connection, which increased the network speed greatly and expand its functions and improved performance of computer to a large extent and allows you to use more networks such as software routers (OpenWRT / DD-WRT / Tomato etc.), firewalls, NAT, network isolation etc.
- 【Large Storage & Flexible Expandability】This Workstation equipped with 64GB LPDDR5-8000MHz + 2TB M.2 2280 PCIe4.0 SSD. There is another PCIe4.0 SSD slot available for up to 8TB, these SSD slots are compatible with RAID0 and RAID1, you can store movies, videos, photos, important files easily. What’s more, it also comes with 1x standard PCIex16 slot(PCIe4.0x4) inside.
How to verify an app’s actual route
Documentation tells you what a vendor says a product can or should do. A claim about what one particular run transmitted needs evidence tied to that run: the product and version, relevant settings, the local model’s readiness, and a trace or log that identifies the destination and, where available, request content. Network activity alone may show a connection without revealing what payload it carried; a request log may expose content but not prove every network destination. Use both kinds of evidence where available, and distinguish observations from documented behavior.
Quick Recap
- Record the exact setup. Note the app name and version, model and whether it is downloaded and ready, account or license edition, privacy settings, logging settings, and whether cloud use is allowed by the organization.
- Identify the route decision. Find the documented conditions for local readiness and fallback. Determine whether a missing or unsupported local model blocks the request, prompts for approval, or can trigger cloud inference.
- Inspect the context boundary. Check whether the app sends only the prompt or also conversation history, open or adjacent files, cursor position, tool output, file types, or framework information. Avoid testing with sensitive material until you understand that boundary.
- Check each destination separately. Distinguish model-provider traffic from downloads, catalog refreshes, product telemetry, organization logging, and any proxy processing. Confirm which settings control each destination.
- Run a controlled check. With approved test content, compare behavior when the local model is ready and when it is unavailable, and repeat with cloud fallback permitted and blocked if the app supports those states. Record route indicators and logs; do not infer payload contents from a connection alone.
- Choose a policy and verify it. If sensitive data must not leave the device, disable cloud routing or select a documented fail-closed mode, then test that an unavailable local model produces an error rather than a cloud request.
Questions to ask before using local-first AI with sensitive material
- What exact condition triggers cloud fallback, and can an administrator disable it?
- Does the request include only typed text, or also conversation history, source snippets, cursor location, tool output, or other context?
- Are model downloads and catalog refreshes separate from inference, and can the app operate from cached resources offline?
- Which telemetry fields contain event metadata, and do any logs store full prompts, responses, or context? Where are those logs sent and who can access them?
- Which provider or intermediary receives each category of data, and what processing-region and retention commitments apply to this edition and configuration?
- Can users inspect a request log or route indicator, and what evidence demonstrates that cloud routing is blocked?
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
