Available software products with local AI usage


1	Description: List of software solutions to run AI or other LLM models on your own, local hardware. Permalink: https://go.xmg.gg/ai-enabled-software						Compatibility with AI hardware acceleration (empty cells mean "unclear" or "tba")									Memory requirements (minimum or recommended)
2	Last update: 2025-03-22 Please also see the additional tab "Notes and Disclaimer" above.						Intel				AMD				NVIDIA	Memory requirements (minimum or recommended)
3	Name	Link	License model	Trial version	Rating (by XMG)	Description	CPU	iGPU	dGPU	NPU	CPU	iGPU	dGPU	NPU	dGPU	System memory (DRAM)	Video memory (VRAM)
4	Amuse AI	Click here	Free to use (Pro version available)	←	▰▰▰▰	Easy to install implementation for Stable Diffusion with an easy to use UI, and direct download of popular models from HuggingFace.com. More user-friendly than AUTOMATIC1111 and other Stable Diffusion web UI solutions.	✅				✅	✅	✅	✅		24 GB	8 GB
5	aTrain (with OpenAI Whisper)	Click here	Free to use	←	▰▰▰▰	Open-source interface for OpenAI's "Whisper" engine, providing extremely reliable speech recognition in a large number of languages. Used for transcription of audio files, such as podcast recordings. Is able to recognize individual speakers. Output as TXT, SRT, JSON and similar open, text-based formats.	✅				✅				✅	16 GB	5 GB
6	Audacity	Click here	Free to use	←	▰▱▱▱	Audio editor. Introduced AI-accelerated effects in early January 2024 as plugins. See blogpost.	✅				✅				-	-	-
7	Blackmagic DaVinci Resolve	Click here	Free to use (Pro version available)	←	▰▰▰▱	Professional video editing software, competitor to Adobe Premiere. Free version supports unlimited projects with resolutions of up to 4K/60Hz. Paid license supports higher resolutions, and a number of AI-features through DaVinci Neural Engine.	✅	✅	✅		✅	✅	✅		✅	16 GB	4 GB
8	Braina	Click here	Subscription- model	✅	▰▰▱▱	Advanced suite for PC voice control and dictation. Old-fashioned interface, but it supports a variety of LLMs under the hood, either run locally or connected to web service API such as GPT4o. Uses OpenAI's "Whisper" for local voice recognition.	✅				✅						-
9	Canvid	Click here	Lifetime license	✅	▰▱▱▱	Video editor with easy-to-use interface for on-screen tutorials. Can use AI to generate audio-track based on your voice training and lip-sync your video footage to audio.	✅				✅
10	Capture One Pro	Click here	Subscription or Lifetime license	✅	▰▰▱▱	Professional photo editor, alternative to Adobe Lightroom. Supports various tools and effects with AI usage, such as object masking. See knowledge base for details.	✅		✅		✅				✅	8 GB	4 GB
11	Cephable	Click here	Free for non-commercial individual use	←	▰▱▱▱	Allows user to program actions based on human input such as gestures, facial expressions, head-pose and voice. Currently targeted at gaming, but actions can be all-purpose in daily computing. Requires webcam and microphone access.	✅				✅
12	Generate by Iterate.ai	Click here	Contact sales	✅	▰▱▱▱	Connect document sources (local storage or 3rd party cloud storage providers) and retrieve information directly from your documents. Like NotebookLM, but runs locally on your PC.	✅				✅				✅
13	GIMP	Click here	Free to use	←	▰▰▱▱	GIMP plugins for upscaling, object masking and image generation through the popular Stable Diffusion models, including Controlnet and other state-of-the-art enhancements. Requires a multi-step installation process, see guide.	✅	✅	✅	✅	✅	-	-	-	✅	16 GB	6 GB
14	Intel AI Playground	Click here	Free to use	←	▰▱▱▱	Chat, image generation and image enhancement. Able to pick between multiple open-source models.	✅	✅	✅	-	✅	-	-	-	-
15	LM Studio	Click here	Free for personal use	←	▰▰▰▰	Run text-based large language models (LLMs) locally through an easy-to-use interface. Supports downloading any open-source model from Hugging Face. Ideal for both beginners and advanced users. See also: Ollama.	✅	✅	✅		✅	✅	✅	✅	✅	16 GB	6 GB
16	Luminar Neo	Click here	Subscription or Lifetime	✅	▰▰▱▱	User-friendly photo and video editor for casual users. Luminar's streamlined user-interface has won the reddot award in 2022. Comes with a high number of AI-supported enhancements. See this article for some examples.	✅				✅				✅	8 GB	4 GB
17	Lykos AI / Stability Matrix	Click here	Open source (AGPL-3.0)	←	▰▰▰▰	Comprehensive platform for managing and running various AI models, including Stable Diffusion. Offers a user-friendly interface for image generation, model management, and workflow creation. Supports both local and cloud-based deployment options for flexibility and scalability.	✅				✅				✅
18	Open Notebook	Click here	Open source (MIT)	←	▰▰▰▱	AI-powered note-taking application that integrates with large language models for enhanced productivity. Features include smart autocomplete, semantic search, and automatic summarization. Allows users to interact with their notes using natural language queries.	✅				✅				✅
19	SiYuan	Click here	Open source (AGPL-3.0)	✅	▰▰▰▱	Open-source personal knowledge management and note-taking software with a focus on privacy and data ownership. Features include block-level referencing, bidirectional link, and graph visualization. Supports local-first storage and end-to-end encryption for cloud sync.	✅				✅				✅
20	SpeedPulse (with OpenAI Whisper)	Click here	Lifetime license	✅	▰▰▰▱	Advanced speech-to-text solution powered by OpenAI's Whisper model. Offers real-time transcription, speaker diarization, and support for multiple languages. Features include custom vocabulary, automated punctuation, and integration with popular video conferencing platforms.	✅				✅				✅
21	Topaz Labs	Click here	Lifetime license	✅	▰▰▰▱	Includes Topaz Video Enhance AI and Topaz Gigapixel to enhance and upscale videos and images. The upscaling algorithsm are very advanced, including custom-based models for different kind of pictures, such as photos, illustrations, schematics etc.	✅	✅	✅		✅	✅	✅		✅	16 GB	4 GB
22	uttertype (with OpenAI Whisper)	Click here	Open source (MIT)	←	▰▰▰▱	Open-source speech-to-text application utilizing OpenAI's Whisper model for accurate transcription. Offers a simple interface for converting audio files to text, supporting multiple languages. Ideal for quick transcription tasks and integrating speech recognition into other applications.	✅				✅				✅
23	VEGAS Pro Edit 365	Click here	Subscription or Lifetime license	✅	▰▰▱▱	Professional video editing, alternative to Adobe Premiere. Starts with a stream-lined, user-friendly interface, but also comes with professional features. See this article for examples.	✅				✅					8 GB	4 GB
24	VNCLagoon AI	Click here	Contact sales	✅	▰▱▱▱	Designed for small to medium businesses. Can interface with 3rd party collaboration tools such as Microsoft Teams to provide meeting minutes and similar insights. Runs entirely locally on your PC.	✅				✅
25	XSplit Broadcaster	Click here	Lifetime license	✅	▰▱▱▱	Can be an alternative to OBS for screen sharing, live streaming and recording. Supports AV1 recording via Intel QuickSync among other features. Can interface with XSplit VCam to enhance camera image.	✅	✅			✅		✅		✅	8 GB	2 GB
26	XSplit VCam	Click here	Lifetime license	✅	▰▱▱▱	Uses AI to enhance your webcam image beyond the built-in capabilities of your collaboration tool (e.g. Teams). Includes features such as denoising, background blur, etc.	✅	✅			✅		✅		✅	4 GB	2 GB


1	Explanation of "Rating" collumn
2	The rating is based on a combination of factors, such as: • how integral and novel is AI to the core function of the software • usefulness in daily life or work • ease of installation and use • price Some rating have been assigned by merely reading about the software, not by actual use. Rating is highly subjective and only offered to avoid sorting alphabeticaly.
3
4	License disclaimer
5	For details on license models, please check with each software vendor. Especially for "free to use" software, double-check if this includes commercial use.
6
7	Notes on GPU and NPU support
8	Support for GPUs and NPUs has been evaluated using available public sources, but information remains limited. Most software vendors do not explicitly specify which hardware accelerators their solutions support. It is evident that cross-vendor compatibility is very challenging, as each hardware vendor requires distinct APIs: • NVIDIA GPUs rely on CUDA • AMD GPUs utilize ROCm • Intel's GPU and NPU solutions require OpenVINO and/or oneAPI Microsoft is working to unify hardware access through DirectML, akin to how DirectX standardized real-time 3D acceleration. • On the NPU side, DirectML already Intel, AMD and Qualcomm. • But on the GPU side, only Intel GPUs are supported. A roadmap for GPU support with AMD or NVIDIA is not yet available in DirectML. DirectML operates within the Windows Subsystem for Linux (WSL) environment, making it potentially more cross-platform compatible than DirectX. However, it will remain to be seen if Microsoft and chipvendors really aim for a true platform-independent, long-term solution.