What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
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
Yes, an older laptop can run useful local-AI workloads, but not every model or task will feel fast. Cherry Studio handles document questions, Jan provides a desktop chatbot, whisper.cpp transcribes audio, and Continue adds local coding help to VS Code. The practical formula is a small model, a focused job and patience for CPU latency.
Can you run a local LLM on an old laptop?
You can, provided you treat local AI as a set of specific workloads rather than a replacement for a fast cloud service. The four projects below target different jobs: document retrieval, general conversation, speech transcription and coding assistance. A slow laptop may handle one of those jobs acceptably while struggling with another.
How-To Geek’s 2026 guidance uses 8 GB of RAM as a minimum and 16 GB as a better target. Those figures are broad advice, not a compatibility rule or a tested guarantee for every project. A machine with 8 GB may need a particularly small model and little else running; 16 GB gives the operating system and the AI runtime more room, but still does not guarantee quick responses.
Some small models in the 4–12 billion-parameter range can run on a CPU, according to the same article. Actual speed and usefulness depend on the laptop, model, runtime, quantization, prompt, document size and task. No controlled benchmark for these four projects on a specified slow-laptop configuration is established, so expect to experiment rather than rely on a promised response time.
#1 Best Overall
- 2K IPS TOUCHSCREEN DISPLAY - 1920 x 1200 resolution delivers incredible detail, wide-viewing angles, and lifelike color reproduction
- AMD RYZEN AI 5 430 PROCESSOR - Unlock powerful AI-driven experiences with a Copilot+ PC powered by an AMD Ryzen AI processor designed to enhance creativity, simplify and streamline your day, and give you valuable time back to do more
- ENJOY UP TO 19 HOURS AND 30 MINUTES OF BATTERY LIFE - HP Fast Charge restores battery from 0 to 50% in approximately 45 minutes
- AMD RADEON 840M GRAPHICS - Built in for thrilling gaming performance, high resolution display support and hardware accelerated encoding with or without a discrete graphics card
- STORAGE AND MEMORY - 512 GB PCIe Gen4 NVMe M.2 SSD offers fast speed and efficient storage; and 16 GB DDR5 RAM memory boosts performance with higher bandwidth
The key distinction is between focused or delayed work and work that demands an immediate answer. Batch transcription can be practical when you can wait. CPU-based code completion or chat may feel noticeably slower than cloud completion, particularly with a larger model. A small, narrow task is usually a better fit than a long, complex request.
Which local AI tool fits your job?
| Project | Best use | How you use it | Slow-laptop fit | Privacy point to check |
|---|---|---|---|---|
| Cherry Studio | Questions over your documents with a knowledge base (RAG) | Desktop client connected to a local inference service such as Ollama or LM Studio | Good for deliberate document queries when you can tolerate waiting | Inference, embeddings and OCR must each be local if the documents must stay on the computer |
| Jan | General-purpose desktop chatbot | Integrated desktop interface aimed at people who do not want command-line setup | Convenient starting point; response speed still depends on the selected model | Confirm where the model and any optional service process your prompts |
| whisper.cpp | Speech-to-text transcription | CPU-oriented transcription workflow; file conversion may require FFmpeg | Batch jobs are more realistic than real-time transcription on a low-power CPU | Keep audio and transcription processing local if confidentiality matters |
| Continue | Code autocomplete and chat inside VS Code | VS Code extension connected to a local model | Use the smallest practical model for autocomplete; reserve larger models for chat | Check the extension, model and any provider configuration before sending code anywhere |
Cherry Studio: ask questions about local documents
Cherry Studio is a desktop client that can connect to local inference services, including Ollama and LM Studio. Its distinctive use here is a knowledge-base workflow: you add documents, let the application index them, and ask questions that retrieve relevant passages before generating an answer. That is more useful than opening a general chatbot when your goal is to work through manuals, notes or other private files.
What the local document workflow involves
- Install Cherry Studio and a local inference service. The client is the interface; the inference service supplies the language model.
- Connect Cherry Studio to that local service and select a model small enough for your laptop’s memory and CPU.
- Create a knowledge base and add the documents you want to search. Large collections and long files increase indexing and query work.
- Choose embedding and OCR options deliberately. Cherry Studio’s current documentation says its built-in embedding and OCR models can run offline after they are downloaded and do not require a provider API.
- Ask a narrowly scoped question first, then inspect whether the retrieved material actually supports the answer before trusting it.
The privacy benefit is conditional. If the language model, embeddings and OCR all run locally, the document-processing path can remain on the laptop. If you select a cloud model or cloud embedding service, document text or images may leave the computer. Cherry Studio’s documentation also describes its lightweight offline embedding and OCR choices as potentially less accurate than stronger cloud alternatives, so privacy and retrieval or recognition quality can involve a trade-off.
When Cherry Studio is a poor fit
A very large archive, image-heavy material or a need for instant answers can overwhelm an older CPU. Start with a small folder and a few representative questions. If indexing consumes too much memory, close other applications, reduce the collection and use a smaller model; do not assume that changing the front-end alone will solve a runtime bottleneck.
Rank #2
- BUILT FOR COLLEGE. AND BEYOND — MacBook Air with the M5 chip packs blazing speed and powerful AI capabilities into an incredibly portable design. And with up to 18 hours of battery life,* this thin and light powerhouse is ready to take on almost any major, just about anywhere.
- TEAR THROUGH TOUGH ASSIGNMENTS — With its faster CPU and unified memory, the M5 chip delivers even more performance and fluidity across apps, making multitasking and creative workflows smooth and responsive. A powerful Neural Engine and next-generation GPU with Neural Accelerators give you a powerful platform for AI.
- MAKE QUICK WORK OF YOUR TO-DO LIST — Apple Intelligence helps you write, express yourself, and get things done effortlessly — whether it’s for school or everyday life. With groundbreaking privacy protections, it gives you peace of mind that no one else can access your data — not even Apple.*
- UP TO 18 HOURS OF BATTERY LIFE — MacBook Air delivers incredible battery life with amazing performance, so you can power through a full day of classes without worrying about plugging in.
- A BRILLIANT 13.6-INCH DISPLAY* — The gorgeous Liquid Retina display on MacBook Air supports 1 billion colors, making photos and videos pop with rich contrast and sharp detail, and text appears supercrisp. So everything — from class presentations to movies to games — looks truly stunning.
Jan: a simpler desktop chatbot
Jan is the approachable option for someone who wants a desktop chat experience without assembling a command-line workflow. Its appeal is convenience: you work in an application rather than configuring every runtime component yourself. The trade-off is less granular control than using a lower-level runtime directly.
A sensible first run
- Install Jan and let it complete its initial model setup.
- Begin with the smallest model that can answer your intended questions. A slow laptop may produce usable output eventually, while a larger model can make every exchange painfully slow or exhaust available memory.
- Test with a short prompt and a modest response length before attempting long documents or multi-step analysis.
- Watch for memory pressure and excessive swapping. If the system becomes unresponsive, stop the generation, close unused programs and move to a smaller model.
Jan is best understood as a convenient local chatbot, not proof that every chat model will run well on old hardware. Verify the model’s source and the application’s provider settings if privacy is important; a desktop interface does not automatically mean every request is processed locally.
whisper.cpp: local speech transcription on the CPU
whisper.cpp is a CPU-oriented speech-to-text project. It is a practical choice when you want to turn recordings into text without sending the audio to an online transcription service. On a low-power laptop, the important decision is whether you need a completed transcript later or words appearing while someone speaks.
The Tool Desk
Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Batch transcription versus real time
- Batch transcription: give the program an existing recording and wait for the output. The source article presents this as the more realistic option for a low-power CPU.
- Real-time transcription: process speech as it happens. CPU latency, audio handling and model workload can make this unreliable or uncomfortable on an older machine.
Basic workflow
- Build or install whisper.cpp for your operating system and confirm that the chosen model is available locally.
- Prepare a short test recording. If the file format is not accepted, use FFmpeg to convert it; some audio workflows need that extra utility.
- Run a batch transcription and measure whether the waiting time is acceptable for your use case.
- Review names, technical terms and punctuation manually. Transcription quality depends on the recording, speakers and model, not only on CPU speed.
No specific microphone, storage device or operating-system combination is established for this project, so compatibility should be checked against your own machine. Keeping the executable, model and audio files local helps with privacy, but you still need to verify that no separate upload or cloud-processing step has been added by your surrounding workflow.
Rank #3
- AN AMAZING MAC AT A SURPRISING PRICE — With an incredibly portable and durable aluminum design, up to 16 hours of battery life,* and the A18 Pro chip, MacBook Neo is ready to go wherever school takes you.
- FOUR STUNNING COLORS. ONE DURABLE DESIGN — Choose from four beautiful colors — Silver, Blush, Citrus, or Indigo — each with a color-coordinated keyboard. And MacBook Neo is made with a durable recycled aluminum enclosure that helps it reach 60 percent recycled content by weight — the most ever in any Apple product.*
- FLY THROUGH EVERYDAY ASSIGNMENTS — Whether you’re cramming for finals, using Apple Intelligence* to summarize class notes, creating presentations, or even playing the latest Apple Arcade game,* MacBook Neo delivers the performance and AI capabilities you need to get things done.
- UP TO 16 HOURS OF BATTERY LIFE — MacBook Neo delivers all day battery life, so you can power through from early morning classes to late night study sessions without worrying about plugging in.
- A VIBRANT 13-INCH DISPLAY* — The gorgeous Liquid Retina display on MacBook Neo supports 1 billion colors, so photos and videos pop and text is crisp for easy reading.
Continue: local coding help in VS Code
Continue is a VS Code extension for local code autocomplete and code chat. It can be useful for small, well-defined tasks such as drafting a regular expression, explaining a short function or suggesting a compact code change. It is less suitable as an expectation of instant, cloud-like completion on an old CPU.
Configure for latency instead of maximum model size
- Install Continue in VS Code and connect it to a local model provider.
- Choose the smallest model that produces acceptable tab-completions. Keep the context you send to the model focused on the current file or function.
- Use the chat panel for tasks that justify waiting, such as explaining an error or proposing a multi-step refactor.
- Test completion and chat separately. A model that is tolerable for occasional chat may still be too slow for every keystroke.
Nick Lewis, the How-To Geek author, recommends: “I’d recommend using the smallest possible model for tab-autocomplete and only using a larger model for the chat panel to get the best performance-to-utility ratio.” That is a performance strategy, not a benchmark result. On a particularly constrained laptop, using the same small model for both may be the more comfortable choice.
Before opening a private repository in chat, inspect the provider configuration and extension settings. A local extension can still be configured to call a remote model, and code context can include more than the snippet you intended to share.
Outdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchWindows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallHow to choose a model and workload for a slow laptop
Start with the memory you actually have
- 8 GB RAM: treat this as the broad minimum cited by How-To Geek, not a guarantee. Keep expectations modest and close unrelated applications.
- 16 GB RAM: the article’s preferred target, giving the runtime more headroom but not ensuring high speed.
- More than 16 GB: extra memory may make larger models or longer context possible, but no project-specific compatibility or speed claim is established here.
Model size matters. The article identifies some 4–12-billion-parameter models as potentially reasonable on a CPU, while emphasizing that results vary by machine, runtime and task. Select for the job: document retrieval needs reliable indexing and answers, transcription needs acceptable recognition, and autocomplete needs low latency. A larger model is not automatically better when every token takes longer.
Rank #4
- PROFESSIONAL PERFORMANCE & MOBILITY - The HP ZBook 8 G1i builds on the legacy of the ZBook Power series, offering pro-level performance in a sleek, mobile design. Built for 3D rendering, simulation, and AI development, its outstanding power efficiency and extended battery life support uninterrupted productivity, while HP Wolf Pro Security (1 year) provides enterprise-grade protection. ISV certifications ensure reliable performance for apps such as SolidWorks, AutoCAD, ANSYS, Revit, and MATLAB
- POWERFUL PERFORMANCE & GRAPHICS - Equipped with the Intel Core Ultra 7 255H Processor (up to 5.1GHz, 16 cores, 16 threads, 24MB L3 cache) and NVIDIA RTX 500 Ada GPU with 4GB GDDR6 dedicated memory, the AI PC delivers desktop-level performance for rendering, AI, and graphics-intensive workloads. Paired with 64GB DDR5 RAM and a 2TB PCIe NVMe M.2 SSD for seamless multitasking and ultra-fast data access
- PROFESSIONAL DISPLAY - The laptop features a 16" WUXGA (1920x1200) Touchscreen with 300-nit brightness and anti-glare technology for vibrant, comfortable viewing. Native multi-display support with up to 8K@60Hz via Thunderbolt 4 and 4K@60Hz via USB-C and HDMI 2.1. Plus, a 5MP IR privacy-shutter webcam delivers secure facial recognition and crisp video calls with Poly Camera Pro, while AI Noise Reduction & Dynamic Voice Leveling ensure clear, professional audio
- RICH CONNECTIVITY OPTIONS - Stay productive with comprehensive connectivity, including 2x Thunderbolt 4, USB-C 3.2 Gen 2x2, USB-A 3.2 Gen 1, Ethernet (RJ-45), HDMI 2.1, and headphone/microphone combo jack. Features Intel Wi-Fi 7 and Bluetooth 5.4 for ultra-fast wireless performance. The built-in fingerprint reader, backlit keyboard, and numeric keypad enhance security, comfort, and everyday usability
- OPERATING SYSTEM - Pre-installed with Microsoft Windows 11 Pro, offering enterprise-grade security with BitLocker and Remote Desktop, designed to support demanding professional applications and enhanced by AI Copilot for smarter, more efficient productivity across business and creative tasks
Match the tool to your tolerance for delay
| If you need… | Prefer… | Accept this limitation |
|---|---|---|
| Answers grounded in a handful of files | Cherry Studio with local inference and local retrieval components | Indexing and responses may take time; lightweight embeddings can be less accurate |
| A familiar chat window | Jan | Convenience replaces some low-level control |
| Transcripts from recordings | whisper.cpp in batch mode | Waiting is preferable to expecting real-time output |
| Occasional coding suggestions | Continue with a small autocomplete model | CPU completion will not feel instantaneous |
Privacy: “local” describes a workflow, not a logo
Installing a local client does not by itself make a workflow private. Identify every component that receives your data: the language model, inference runtime, embeddings, OCR, transcription process and any extension provider. Cherry Studio’s offline embedding and OCR documentation is a useful example of why this matters: those components can stay local after download, but stronger cloud services may offer better retrieval or recognition.
For sensitive documents or source code, check provider endpoints before importing data, disable optional cloud features when appropriate and run a small test with network access restricted if your security policy allows it. A local model can also produce incorrect answers, so privacy does not remove the need to verify output.
Troubleshooting common slow-laptop problems
Generation is extremely slow
Switch to a smaller model, shorten the prompt and close memory-heavy applications. Use delayed workflows such as batch transcription or occasional chat rather than continuous autocomplete. If the laptop remains unusable, the workload exceeds what its CPU and memory can comfortably provide.
The system runs out of memory
Stop the current generation, unload the model, reduce context and choose a smaller model. Do not treat 8 GB as permission to run any 4–12-billion-parameter model; the operating system and application overhead also need memory.
Best Value
- FAST RUNS IN THE FAMILY — The 14-inch MacBook Pro with the M5 Pro or M5 Max chip brings next-generation speed and powerful on-device AI to personal, professional, and creative tasks. With all-day battery life, double the starting storage,* and a breathtaking Liquid Retina XDR display, it’s pro in every way.*
- BUCKLE UP — Along with a next-generation CPU, faster unified memory, and up to 2x faster SSD storage,* M5 Pro and M5 Max feature a more powerful GPU with a Neural Accelerator built into each core, delivering faster AI performance and on-device training capabilities. So you can blaze through demanding workloads at mind-bending speeds.
- BUILT FOR AI — Apple silicon, and every major component that powers it, is designed to run demanding on-device AI workloads like LLM inference and training. And Apple Intelligence helps you write, express yourself, and get things done effortlessly with groundbreaking privacy protections at every step.*
- ALL-DAY BATTERY LIFE — MacBook Pro delivers the same exceptional performance whether it’s running on battery or plugged in.*
- MACOS RUNS APPS FAST — All your go-to apps run lightning fast in macOS, including built-in apps like FaceTime and Messages. Plus, built-in virus protection and free software updates help keep your Mac running smoothly and securely.
Document answers are inaccurate
Reduce the knowledge base to relevant files, check that OCR succeeded and confirm that the embedding model is local and appropriate. Cherry Studio notes that its lightweight offline choices can be less precise than stronger cloud alternatives, so accuracy may be the cost of staying offline.
An audio file will not transcribe
Convert it with FFmpeg to a supported format, retry with a short sample and use batch processing. If real-time capture fails, that may reflect CPU or audio-pipeline limits rather than a transcription-model defect.
Autocomplete interrupts your work
Use a smaller model, limit the code context and move complex requests to Continue’s chat panel. Nick Lewis’s small-autocomplete/larger-chat split is intended to preserve responsiveness where it matters most.
Do these 3 things before closing this tab:
1Scan for outdated or missing drivers - takes under a minute2Repair Windows errors before they cause bigger problems3Fix the driver behind crashes, sound loss and screen glitchesWhat you need before installing
- A laptop with enough free RAM for the operating system, application and model; 8 GB is broad minimum guidance and 16 GB is preferable.
- Storage and network access sufficient to download the applications and model files. Exact storage requirements are not established for these four projects.
- For Cherry Studio, a local inference service such as Ollama or LM Studio if you want that architecture.
- For Continue, VS Code and a provider configuration that you have checked for local processing.
- For whisper.cpp, audio files and possibly FFmpeg for conversion. No particular microphone model is required by the evidence available here.
You do not need to buy a dedicated AI computer to try these projects, and no specific hardware upgrade is established as necessary. The sensible approach is to start with one narrowly defined task, use the smallest practical model, and judge the waiting time against the value of keeping processing on your own machine.
Quick Recap
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.
