PromptQuorum: Your Discovery Engine for Open-Weight & Open-Source AI
Directory, Guides & Comparison Tools for Local AI & LLMs
(in plain terms: running an AI model for free on your own computer, with your data staying local)
The State of Local AI · Report 01 · Q4 2026
Few Chickens and Even Fewer Eggs
Very few computers can run a good local AI model, and even fewer apps make it worth the effort. Nine hands-on tests, sourced figures and a claim register show where local AI stands today.
Just published
Added in the last 3 days
- NewReview
Stirling PDF Review: Open-Core PDF Toolkit for Desktop and Self-Hosting
Factual review of Stirling PDF, an open-core PDF editor and API that runs on desktop, Docker or Kubernetes. Install options, license limits and open questions.
November 16, 2026
- NewReview
Docling Review: Open-Source PDF and Document Parser for RAG
Docling review: PDF parser with MIT-licensed code, OCR, table extraction, bounding boxes and an MCP option. pip install docling, Python 3.10+.
November 15, 2026
- NewReview
Marker Review: PDF-to-Markdown Converter with OCR and Table Extraction
Marker review: PDF-to-Markdown converter with Surya OCR and table extraction. Code is Apache-2.0; model weights have a separate licence. Python 3.10+.
November 15, 2026
- NewReview
MinerU Review: Source-Available Document Parser for Markdown, LaTeX and RAG
MinerU review: document parser under a custom Apache 2.0-based licence with OCR, cross-page tables and GPU support. pip install mineru, Python 3.11–3.14.
November 15, 2026
- NewReview
OpenDataLoader PDF Review: Open-Source PDF Parser for RAG
OpenDataLoader PDF review: Apache 2.0 PDF parser for RAG. Needs Java 11+; default mode runs locally per the README. Benchmarks are the vendor's.
November 15, 2026
- NewReview
AI Notebook Review: Free Open-Source NotebookLM Alternative
AI Notebook review: free MIT-licensed Mac and Windows notebook that answers over your documents with citations and can run on Ollama. Not hands-on tested.
November 14, 2026
- Apps tracked
- 370
- Languages
- 9
- Apps verified by their makers
- 24
- Articles published
- 961
- Apps tested hands-on
- 11
- AI assistant calls (MCP)
- 2,946
- Ask Hans answers
- 516
Open directory →
The makers confirm the information presented about their apps.
PromptQuorum has installed and tested the software itself.
See how AI assistants use this site →
See how Ask Hans is used →
Find What You Need
Pick what you're after, then search.
Largest local AI & LLM app directory outside GitHub with real reviews and hands-on testsBrowse 370 apps →Compare Tools
Pick a category, choose up to three tools, and see them side by side. Copy the table or open the full guide.
Pick a category above to start comparing.
Optimize one prompt. Ask 25+ AIs. Compare the answers.
PromptQuorum structures your prompt, sends it to many models at once, and shows where they agree.
- Automatic prompt optimization
- One prompt to 25+ AI models
- Side-by-side comparison
- Consensus scoring
- Local AI & LLM support
Ask your AI about anything on PromptQuorum. Answers straight from the site.
Connect Claude, ChatGPT, or any MCP client to all of PromptQuorum: the app directory, guides, hands-on tests, and the latest updates. Get sourced answers right inside your chat.
Search the whole site
Find guides, reviews, and hands-on tests on any topic, with links to the source.
Compare & choose apps
Side-by-side comparisons, best-fit picks, local alternatives, and plain-language license explanations.
Hardware & VRAM checks
Tell it your GPU, RAM, or Mac and see which apps and models will actually run.
Always up to date
Ask what is new: the latest articles and updates, plus live usage stats of the MCP server.
Try asking
- “I have a MacBook with 16 GB of RAM. Which local chat apps and models will run well on it?”
- “Compare LM Studio and Ollama for a beginner, and check whether either license allows commercial use.”
- “What did PromptQuorum publish recently on prompt engineering? Summarize the most useful article for me.”
Local AI Tools
Three ways to match your hardware to the right software.
Latest Posts
- New
Stirling PDF Review: Open-Core PDF Toolkit for Desktop and Self-Hosting
Factual review of Stirling PDF, an open-core PDF editor and API that runs on desktop, Docker or Kubernetes. Install options, license limits and open questions.
November 16, 2026
- New
OpenDataLoader PDF Review: Open-Source PDF Parser for RAG
OpenDataLoader PDF review: Apache 2.0 PDF parser for RAG. Needs Java 11+; default mode runs locally per the README. Benchmarks are the vendor's.
November 15, 2026
- New
Docling Review: Open-Source PDF and Document Parser for RAG
Docling review: PDF parser with MIT-licensed code, OCR, table extraction, bounding boxes and an MCP option. pip install docling, Python 3.10+.
November 15, 2026
- New
Marker Review: PDF-to-Markdown Converter with OCR and Table Extraction
Marker review: PDF-to-Markdown converter with Surya OCR and table extraction. Code is Apache-2.0; model weights have a separate licence. Python 3.10+.
November 15, 2026
- New
MinerU Review: Source-Available Document Parser for Markdown, LaTeX and RAG
MinerU review: document parser under a custom Apache 2.0-based licence with OCR, cross-page tables and GPU support. pip install mineru, Python 3.11–3.14.
November 15, 2026
- New
AI Notebook Review: Free Open-Source NotebookLM Alternative
AI Notebook review: free MIT-licensed Mac and Windows notebook that answers over your documents with citations and can run on Ollama. Not hands-on tested.
November 14, 2026
- New
OpenMausBot Review: Is This Open-Source Grok Bot Local AI?
OpenMausBot review: open-source Grok Bot alternative on Claude, Codex or Grok agents. Not local by default; Ollama and LM Studio can connect. Free.
November 14, 2026
- New
BrowserOS Review: Open-Source AI Browser With Local Ollama and LM Studio Support
BrowserOS review: a free, open-source AI browser for Mac, Windows and Linux that works with local Ollama or LM Studio models. Privacy, licence, risks.
November 14, 2026
- New
GMKtec EVO-X5 Pro vs DGX Spark: Specs, Price and Verdict
GMKtec EVO-X5 Pro (192 GB, Ryzen AI Max+ PRO 495) against the NVIDIA DGX Spark: sourced specs, US prices, networking, and who should skip each.
November 13, 2026
- New
Local AI on Android in 2026: What Is Possible, What Is Not, and the Apps We Cover
Local AI on Android in 2026: what runs offline, what does not, RAM and chip limits, and the apps we cover by chat, voice, vision, documents and more.
November 13, 2026
Latest Apps Added
- New
Stirling PDF
Open-core PDF toolkit that edits, converts, OCRs, redacts and batch-processes PDFs on the desktop, on a self-hosted server or through an API. — OCR & document parsing
November 16, 2026
- New
OpenDataLoader PDF
Open-source PDF parser that turns PDFs into structured JSON, Markdown and HTML with bounding boxes for RAG and LLM pipelines, runs locally and also auto-tags PDFs for accessibility — OCR & document parsing
November 15, 2026
- New
Docling
Open-source document parser started by IBM Research and hosted by the LF AI & Data Foundation that converts PDF, DOCX, PPTX and more into structured Markdown, JSON and HTML with bounding boxes and OCR; parses on your machine, with MCP via the separate docling-mcp package — OCR & document parsing
November 15, 2026
- New
Marker
Open-source PDF-to-Markdown converter that also handles DOCX, PPTX, EPUB and images; outputs Markdown, JSON and HTML with table and layout preservation; runs locally or via the hosted Datalab API — OCR & document parsing
November 15, 2026
- New
MinerU
Document parser with public source code that converts PDF, DOCX, PPTX, EPUB, HTML and images into Markdown, HTML and LaTeX with cross-page table merging and OCR; supports GPU acceleration on NVIDIA, AMD, Intel and Apple Silicon per the README — OCR & document parsing
November 15, 2026
- New
BrowserOS
Free open-source Chromium AI browser for Mac, Windows and Linux whose Chat Mode can use local Ollama or LM Studio models or your own cloud provider keys — Browser sidebars & AI browsers
November 14, 2026
- New
AI Notebook
Free, MIT-licensed notebook for Mac and Windows that answers questions over your own documents with citations and can keep every AI call on your device through Ollama — Document & PDF chat
November 14, 2026
- New
Obscura
Open-source headless browser engine written in Rust that gives AI agents and scrapers a Chrome DevTools Protocol endpoint, usable as a Puppeteer and Playwright drop-in — Web scraping for AI
November 14, 2026
- New
OpenMausBot
Open-source desktop chat app that runs a team of AI bots on the Claude, Codex or Grok agents you already have, with an option for local models — Multi-agent orchestration
November 14, 2026
- New
Phantasm
Free iPhone chat app for your own Ollama, vLLM or llama.cpp server, with no account and an optional companion server for web search and image generation — General chat clients
November 13, 2026
Latest Founder/Maker Reviews
StratoSort
Ben Kennedy, Ph.D., maker of StratoSort — I built StratoSort for people whose files are too private to upload to a cloud AI service.
November 10, 2026
Sprich
Jakob, co-founder — We built Sprich because dictating in German on the Mac was either poor or sent everything to a US cloud.
November 10, 2026
Atlarix
Amariah Abishai, Founder, Norah Labs — Built Atlarix to be an AI coding agent trustworthy with a real codebase: one that plans, edits, runs the project's own tests, and never changes a file until the diff is approved — with Ollama or LM Studio your prompts and code go only to that local model.
November 7, 2026
Sablate
Furkan Bağcı, developer at FuumiSoft — Sablate exists to do the whole subtitle job on the computer where the footage already is: Whisper transcription, translation, editing and burn-in in one window, with no upload and no per-minute fee.
November 7, 2026
RAMen
November 6, 2026
GreenCube
Hector Gras, founder — I'm 16, and I started GreenCube because I think AI is meant to be owned, not rented.
November 6, 2026
TecAdRiseBot
Chris, TecAdRise, maker of TecAdRiseBot — We built TecAdRise Bot because we needed a local multi-agent desktop shell on Windows that we could own end to end, without locking the workflow into a hosted chat UI.
October 28, 2026
Lyonesse
Josip Markovic, developer — I built Lyonesse because I wanted meeting notes without a bot joining the call and without my audio going to someone's server.
October 26, 2026
Pajdo: Health & Fitness Buddy
Josip Markovic, developer — I wanted my Apple Watch data explained like a friend would, without sending my health data anywhere.
October 26, 2026
Loci AI
Michael Waldman, Founder & CEO, Loci AI, Inc. — We built Loci because we believe capable AI should be a public good, not a subscription that sends every word to a data center. The biggest surprise of the last year is how far open models have come. On a Mac, Loci now runs models that score above frontier cloud models on independent benchmarks, for free and fully offline. Most of our work goes into making that feel effortless: picking the right model for your hardware, keeping long sessions stable, and letting your phone use your desktop's power. We're particularly excited about our universal harness, which is focused on increasing performance of smaller models. Most importantly, Loci is working on an inference engine, which gives us full control of local execution and will continue to push the frontier of local AI.
October 23, 2026
Latest Hands-on Tests
- New
Univocal on an M4 Pro Mac with 24 GB: Dictation Worked in Three Languages, Read Aloud Did Not Start
A first session with a Mac app that dictates and reads aloud: a careful first run with licence terms up front, dictation in English, French and German, a premium-voice route the tester could not follow, and a read-aloud function that did not start when used.
November 13, 2026
Sprich on an M4 Pro Mac with 24 GB: Local Dictation That Types Into Any App
Two sessions, 9 and 13 November 2026: a polished five-step setup, quick dictation in English and German once the tester used the key correctly, a meeting recorded and split into two speakers despite poor room sound, a German-only interface, and several side features not tried.
November 10, 2026
Locally Uncensored (LU) on a 24 GB Mac: Chat, Document Chat, Images and a Coding Agent in One App
A first session on an Apple M4 Mac with 24 GB: a considerate setup and a wide feature set, with PDF, Word and Excel files not read.
October 24, 2026
Caelum on an Android Phone: Reads a Parcel Slip, Stumbles on Facts and Logic (Other AIs Slipped on the Logic Too)
The first PromptQuorum hands-on test run on a phone instead of a Mac. A OnePlus 8 Pro with 12 GB, a 1,020 Mo model download and no word on which model it is. The aim was simply to see how the app works and what is possible; setup caused no problems.
October 21, 2026
LTX Desktop on a 24 GB Mac: Stopped at Setup, Below the Vendor's Memory Floor
Hands-on test on a 24 GB Mac. The app launched smoothly, then the setup screen asked for 129.1 GB of model downloads and the tester gave up: 24 GB of memory is not enough for this app, he says.
October 20, 2026
Msty Studio on a 24 GB Mac: It Found My Ollama Models at Once, What It Is For Stayed Unclear to Me
Hands-on test on an M3 Mac with 24 GB. Existing Ollama models appeared without a second download. No image was generated.
October 16, 2026
Unsloth on a 24 GB Mac: Polished App, Out of Memory Before the First Picture
Hands-on test on a 24 GB Mac, aimed at picture generation. Chat worked; picture models were refused for lack of memory, and no picture was made.
October 16, 2026
DiffusionBee on a 24 GB Mac: Generating Worked at Once, Editing Mostly Did Not
Hands-on test on an M3 Mac with 24 GB. Pictures came out in about a minute; targeted edits to an existing picture mostly failed.
October 16, 2026
Draw Things on a 24 GB Mac: Fast When It Works, Blind Until It Does
Hands-on test on a MacBook Pro with M3 Pro. The first prompt returned a gray square. The second model worked.
October 9, 2026
BoBe Alpha on a 24 GB Mac: Easy to Install, Restarted When Deleted
Hands-on test on a MacBook Pro with M3 Pro. The in-app assistant did not name the macOS setting it needs, and the app restarted itself when the tester tried to delete it.
October 9, 2026
Trending / Most-Read
Snapshot as of October 4, 2026 (last 28 days)
- Uncensored Local LLMs for Creative Writing
4,218 clicks
- 無検閲ローカルLLMと創作活動
1,918 clicks
- 无审查本地LLM与创意写作
1,194 clicks
- 로컬 LLM 순위·추천: VRAM 등급별 비교
790 clicks
- Best Local LLM Apps for Android
679 clicks
- Small Local LLM Models
648 clicks
- Fastest Local LLMs for Low-End PCs
570 clicks
- Beste lokale LLMs im Vergleich
563 clicks
- Best CPU-Only LLM
552 clicks
- Apple MLX vs. NVIDIA CUDA for Local LLMs
548 clicks
- Best AMD Mini PC for Local LLMs
491 clicks
- iPhoneで使えるLocal LLMアプリ比較
458 clicks
- Best Local LLMs by VRAM Tier
424 clicks
- コーディング用ローカルLLM:Kimi K2.6 vs Qwen vs Devstral
421 clicks
- Local LLM Hardware Guide
406 clicks
- 창의적 글쓰기를 위한 무검열 로컬 LLM
370 clicks
- Apple Silicon M5 for Local LLMs
364 clicks
- 로컬 AI와 로컬 LLM이란?
348 clicks
- Best Local LLMs for Coding
334 clicks
- Best AMD GPUs for Local LLMs
327 clicks
Recently Updated Content
- Updated
Local LLM Security and Privacy Checklist: 12 Steps to a Safe Setup
Updated November 15, 2026
- Updated
Best GPUs for Local LLMs in 2026: VRAM Tiers, Speed and Prices
Updated November 15, 2026
- Updated
How to Run Qwen 3.6 and Qwen 3.8 Locally in 2026: Ollama and LM Studio Setup Guide
Updated November 15, 2026
- Updated
Local LLM Apps & Software Directory: 370 Tools for Running AI Locally (2026)
Updated November 15, 2026
- Updated
Download & Install Ollama on macOS, Windows & Linux
Updated November 14, 2026
- Updated
Install LM Studio: GUI Setup for macOS, Windows & Linux
Updated November 14, 2026
- Updated
Open WebUI in 2026: The Self-Hosted ChatGPT Interface for Ollama and Any AI Model
Updated November 14, 2026
- Updated
SillyTavern in 2026: The Free Frontend for AI Character Roleplay and Long-Form Chat
Updated November 14, 2026
- Updated
Local LLM Hardware: GPU vs Mini PC vs Mac Compared
Updated November 14, 2026
- Updated
Best AMD Mini PC for Local LLM 2026: AOOSTAR, Minisforum, Beelink, GMKtec Compared
Updated November 14, 2026
Recent Mentions
Most Recent
- Minisforum um890 pro META Preisvergleich — META Preisvergleich
October 4, 2026 · DR 36
- IFA 2026: AMD के लोकल AI PC में लैपटॉप से 192GB मेमोरी वाले NAS तक विकल्प — Studio Global AI
October 4, 2026 · DR 29
- OpenAI được cho là mua hàng chục nghìn máy Mac: Không phải để thay thế các cụm GPU Nvidia — Studio Global AI
October 4, 2026 · DR 29
- ทางสายกลางของ AI ยุคใหม่: ใช้โมเดลจีน Self-Host เพื่อควบคุมข้อมูล หรือเช่า API สหรัฐฯ เพื่อความง่าย — Studio Global AI
October 4, 2026 · DR 29
- Локальные LLM на Android: что и зачем реально можно гонять на телефоне — Hermes Agent
October 4, 2026 · DR 21
- Home Assistant 完整教學 2026:9 萬星開源智慧家庭中樞,HAOS/Docker 安裝、Zigbee 與 Matter 設定 — Dennis 技術視野洞察
October 4, 2026 · DR 11
- Muss ich wirklich Prompt Engineering lernen? — KI einfach nutzen — better-computing.com
October 4, 2026 · DR 6
- Workstation AI vs DGX Spark — Supercat.dev
October 4, 2026 · DR 0
- AnythingLLM — Grokipedia article — Grokipedia
October 2, 2026 · DR 79
- 192GB统一内存加持,消费级硬件单台能跑2840亿参数大模型 — SMZDM
October 2, 2026 · DR 79
What's your biggest local LLM headache?
Are you a founder? Tell us about your app →Did You Know?
AutoGPT is really two differently-licensed projects sharing one repo: the MIT-licensed classic CLI agent, and a separately-licensed hosted Platform product.
Read more: AutoGPT (Classic, Local) Review →