RUNLOCALAIv38
→WILL IT RUNBEST GPUCOMPARETROUBLESHOOTSTARTPULSEMODELSHARDWARETOOLSBENCH
RUNLOCALAI

Operator-grade instrument for local-AI hardware intelligence. Hand-written verdicts. Real benchmarks. Reproducible commands.

OP·Fredoline Eruo
DIR
  • Models
  • Hardware
  • Tools
  • Benchmarks
  • Will it run?
GUIDES
  • Best GPU
  • Best laptop
  • Best Mac
  • Best used GPU
  • Best budget GPU
  • Best GPU for Ollama
  • Best GPU for SD
  • AI PC build $2K
  • CUDA vs ROCm
  • 16 vs 24 GB
  • Compare hardware
  • Custom compare
REF
  • Systems
  • Ecosystem maps
  • Pillar guides
  • Methodology
  • Glossary
  • Errors KB
  • Troubleshooting
  • Resources
  • Public API
EDITOR
  • About
  • About the author
  • Changelog
  • Latest
  • Updates
  • Submit benchmark
  • Send feedback
  • Trust
  • Editorial policy
  • How we make money
  • Contact
LEGAL
  • Privacy
  • Terms
  • Sitemap
MAIL · MONTHLY DIGEST
Get monthly local AI changes
Monthly recap. No spam.
DISCLOSURE

Some links on this site are affiliate links (Amazon Associates and other first-class retailers). When you buy through them, we earn a small commission at no extra cost to you. Affiliate links do not influence our verdicts — there are cards we rate highly that we don't have affiliate relationships with, and cards that sell well that we refuse to recommend. Read more →

SYS · ONLINEUPTIME · 100%2026 · operator-owned
RUNLOCALAI · v38
← Home·Runtimes (/tools)

>Apps directory

What plugs into your local AI runtime. 37 curated apps across 12 categories — chat UIs, coding agents, RAG pipelines, voice, image, browser extensions, editor plugins, mobile + desktop, agent frameworks, productivity, SDK wrappers.

Each entry carries an honest editorial verdict — pros, cons, the runtime it works against, the minimum VRAM, and the privacy posture. Filter to your stack, jump to the detail page, ship.

Filter the directory

URL updates as you change filters — share or bookmark a result. All filters are server-rendered, so the page works without JS.

§ Category
§ Sort by

37 of 37 apps

Open WebUI

Chat UI
★ 4.8
Fully offline

The default chat UI for solo Ollama users. Multi-model, built-in RAG, web search, Docker-friendly.

“Best default chat UI for solo Ollama users. Pick this first; switch only if you outgrow it.”

ollamaopenai-compat
Free· 4GB+ VRAM
★ 56.0k

LangChain

Agent framework
★ 4.0
Hybrid (offline or cloud)

The default agent framework. Pipelines, retrievers, tool-calling — works against any local backend.

“The default agent framework. Heavy on abstractions, deep ecosystem — pick this if you want defaults.”

ollamallama-cppvllmopenai-compat
Free
★ 95.0k

Aider

Coding agent
★ 4.7
Fully offline

Terminal coding agent that edits files via your local model. Git-aware, surgical, fast.

“Best terminal-native coding agent for local models. Qwen 2.5 Coder 32B is its sweet spot.”

ollamaopenai-compat
Free· 24GB+ VRAM
★ 27.0k

LM Studio

Desktop app
★ 4.5
Fully offline

Desktop app that bundles model download + chat + OpenAI-compatible local server. Closed-source but free.

“Best 'first install' desktop app for newcomers. Closed-source but the easiest first-run experience.”

llama-cppmlx
Free tier· 4GB+ VRAM

Cline

Coding agent
★ 4.6
Hybrid (offline or cloud)

VS Code extension that runs a full agent loop locally — reads, writes, runs commands, asks first.

“Best IDE-integrated agent that fully respects 'all local' as a first-class option.”

ollamaopenai-compatanthropicopenai
Free· 24GB+ VRAM
★ 30.0k

Jan

Chat UI
★ 4.5
Fully offline

Privacy-first desktop chat with a curated model catalog. Llama / Mistral / Qwen one click from the app.

“Best one-binary desktop chat. Curated catalog removes 'which model?' decision paralysis.”

llama-cppollamaopenai-compat
Free· 4GB+ VRAM
★ 28.0k

AnythingLLM

Chat UI
★ 4.5
Hybrid (offline or cloud)

Docs-aware chat with workspaces. Drop a folder of PDFs, get a working RAG chatbot in 5 minutes.

“Best fast-RAG app. Workspace model is the right abstraction for doc-corpora chat.”

ollamalm-studioopenai-compat
Free· 8GB+ VRAM
★ 34.0k

LlamaIndex

Agent framework
★ 4.4
Hybrid (offline or cloud)

RAG-first agent framework. Better defaults than LangChain for doc-corpora work; same local-runtime story.

“Best agent framework for RAG-first workloads. Less abstraction than LangChain.”

ollamallama-cppopenai-compat
Free
★ 38.0k

Continue

Coding agent
★ 4.5
Hybrid (offline or cloud)

Open-source autocomplete + chat for VS Code and JetBrains. Local-model-first.

“Best Copilot replacement that defaults to local. Configurable; pair with Qwen 2.5 Coder.”

ollamaopenai-compatanthropicopenai
Free· 12GB+ VRAM
★ 21.0k

PrivateGPT

RAG app
★ 4.4
Fully offline

Air-gappable RAG over your docs. The OG offline-RAG project, now mature and team-friendly.

“Best when air-gap compliance is the requirement. Less polished than AnythingLLM, more configurable.”

ollamallama-cppopenai-compat
Free· 8GB+ VRAM
★ 54.0k

LiteLLM

SDK / proxy
★ 4.5
Hybrid (offline or cloud)

Drop-in OpenAI-compatible proxy across 100+ providers. Route to local Ollama or cloud, same code.

“Best universal LLM proxy. Foundational layer for multi-provider deployments.”

ollamallama-cppopenai-compatanthropic+2
Free tier
★ 15.0k

Draw Things

Image gen UI
★ 4.6
Fully offline

Free, native macOS / iOS Stable Diffusion app. Runs SD3, Flux on a phone (yes, really).

“Best mobile + macOS SD app. Free, native, no Python — runs Flux on Apple Silicon impressively well.”

mlxcoreml
Free· 8GB+ VRAM

GPT4All

Desktop app
★ 4.0
Fully offline

Nomic's free desktop AI with model catalog + chat + Python SDK. Long-standing, open-source.

“Best fully-open-source desktop AI bundler. Less polished than LM Studio, fully MIT.”

llama-cpp
Free· 4GB+ VRAM
★ 70.0k

LibreChat

Chat UI
★ 4.4
Hybrid (offline or cloud)

Open-source clone of the ChatGPT UI with multi-provider routing. Local + cloud in one interface.

“Best if you mix local + cloud models in the same workflow. Strong team features.”

ollamaopenai-compatanthropicgemini
Free· 4GB+ VRAM
★ 22.0k

Ollama Python SDK

SDK / proxy
★ 4.5
Fully offline

Official Python SDK for Ollama. Async, streaming, typed — the right primitive for scripts.

“Foundational primitive for Python scripts against Ollama. Official, maintained, typed.”

ollama
Free
★ 5.5k

Tabby

Coding agent
★ 4.4
Fully offline

Self-hosted coding agent server with team SSO, audit logs, and dashboards. Enterprise-grade.

“Best self-hosted server for teams. SSO + audit logs make it the IT-friendly pick.”

llama-cppopenai-compat
Freemium· 16GB+ VRAM
★ 23.0k

CrewAI

Agent framework
★ 4.3
Hybrid (offline or cloud)

Production-leaning multi-agent framework. Role + goal + task — opinionated and ergonomic.

“Best ergonomic multi-agent framework. Picks defaults you'd otherwise have to argue about.”

ollamaopenai-compat
Free
★ 22.0k

Obsidian Copilot

Editor plugin
★ 4.4
Hybrid (offline or cloud)

Obsidian plugin that wires Ollama / OpenAI into your notes. Inline chat, summarize, prompt-templates.

“Best Obsidian plugin for local LLM in your notes. Pair with Smart Connections for RAG.”

ollamaopenai-compat
Free· 4GB+ VRAM
★ 3.0k

Khoj

RAG app
★ 4.3
Hybrid (offline or cloud)

Self-hosted AI assistant for your notes, emails, docs. Web + mobile + desktop, all local-first.

“Best 'AI second brain' app. Self-hosted, local-first, works against Obsidian.”

ollamallama-cppopenai-compatanthropic
Freemium· 8GB+ VRAM
★ 18.0k

Enchanted

Mobile app
★ 4.5
Fully offline

Native iOS / macOS Ollama client. Beautiful SwiftUI, talks to your home Ollama server.

“Best mobile Ollama client. Native SwiftUI; works against your home Ollama server.”

ollama
Free· 4GB+ VRAM
★ 6.0k

AutoGen

Agent framework
★ 4.2
Hybrid (offline or cloud)

Microsoft's multi-agent framework. Conversation-first orchestration of role-played agents.

“Best for multi-agent role-played workflows. Niche; not the default agent framework.”

ollamaopenai-compat
Free
★ 36.0k

MacWhisper

Voice / transcription
★ 4.6
Fully offline

Native macOS app for Whisper transcription. Drag a file in, get a transcript out.

“Best Whisper desktop app on macOS. Pay once, transcribe locally forever.”

whisper-cpp
Paid· 2GB+ VRAM

Smart Connections (Obsidian)

Editor plugin
★ 4.4
Fully offline

Local semantic search across all your Obsidian notes. Embed-once, query-fast, fully offline.

“Best local semantic search for personal notes. Foundational layer for Obsidian RAG.”

ollamaopenai-compat
Freemium· 4GB+ VRAM
★ 3.0k

Buzz

Voice / transcription
★ 4.2
Fully offline

Open-source Whisper transcription with mic + file modes. Cross-platform Qt app.

“Best open-source Whisper desktop app. Cross-platform, free, less polish than MacWhisper.”

whisper-cpp
Free· 2GB+ VRAM
★ 13.0k

Verba

RAG app
★ 4.2
Hybrid (offline or cloud)

Weaviate's open-source RAG demo turned production. Strong defaults, opinionated stack.

“Best for 'don't make me choose chunking strategy' teams. Opinionated stack works.”

ollamaopenai-compatanthropic
Free· 8GB+ VRAM
★ 7.5k

Krita AI Diffusion

Image gen UI
★ 4.3
Fully offline

Krita plugin that wires ComfyUI into a real digital-art workflow. Inpaint, outpaint, upscale.

“Best 'SD as digital-art tool' integration. Real Krita workflow, not a wrapper UI.”

comfyui
Free· 8GB+ VRAM
★ 7.5k

Ollama JS / TS SDK

SDK / proxy
★ 4.4
Fully offline

Official Node + browser SDK for Ollama. ESM-first, typed, streaming.

“Foundational primitive for Node + browser apps against Ollama. ESM-native, typed.”

ollama
Free
★ 4.5k

Page Assist

Browser extension
★ 4.4
Fully offline

Browser sidebar that talks to your local Ollama. Summarize pages, chat, vision support.

“Best 'sidebar AI' browser extension that's truly local-first.”

ollamaopenai-compat
Free· 4GB+ VRAM
★ 5.0k

Msty

Desktop app
★ 4.3
Hybrid (offline or cloud)

Desktop app focused on side-by-side multi-model chat. Compare local vs cloud answers in one view.

“Best 'compare local vs cloud answers' workflow. Niche but well-designed.”

ollamaopenai-compatanthropicopenai+1
Free tier· 4GB+ VRAM

Reor

Productivity
★ 4.1
Fully offline

AI note-taking app that builds connections between your notes automatically. Local, open-source.

“Best AI-first note app that's actually local. Niche but well-executed.”

ollamaopenai-compat
Free· 4GB+ VRAM
★ 8.0k

Twinny

Coding agent
★ 4.2
Fully offline

Free, lightweight VS Code copilot that runs entirely on Ollama. Strong on autocomplete.

“Best minimal-surface Copilot-replacement that's been Ollama-native since day one.”

ollama
Free· 8GB+ VRAM
★ 3.5k

Diffusion Bee

Image gen UI
★ 4.0
Fully offline

One-click Stable Diffusion app for macOS. No setup, just run.

“Easiest macOS SD app — picks defaults so you don't have to.”

coreml
Free· 8GB+ VRAM
★ 13.0k

OpenedAI-Speech

Voice / transcription
★ 4.2
Fully offline

Drop-in OpenAI TTS-compatible server. Self-hosted, talks to local voice models.

“Best 'drop-in local TTS for OpenAI clients'. Bridge solution for existing pipelines.”

openai-compat
Free· 4GB+ VRAM
★ 1.5k

Leo (Brave)

Browser extension
★ 4.2
Hybrid (offline or cloud)

Brave's built-in AI assistant. Configurable to talk to local Ollama out of the box.

“Best built-in browser AI for Brave users. Local mode is a checkbox, not a hack.”

ollamaopenai-compat
Free· 4GB+ VRAM
★ 18.0k

Maid

Mobile app
★ 4.0
Fully offline

Android Ollama client + on-device fallback for small models. Cross-platform Flutter.

“Best cross-platform Android-friendly Ollama client. Falls back to on-device for tiny models.”

ollamallama-cpp
Free· 2GB+ VRAM
★ 2.0k

Codeium (with local backend)

Editor plugin
★ 4.0
Fully offline

Codeium self-hosted enterprise backend lets the popular IDE plugin run fully on your hardware.

“Best 'enterprise Copilot' replacement when self-hosting is mandatory. Paid tier.”

custom
Paid· 24GB+ VRAM

Khoj CLI

Productivity
★ 4.2
Fully offline

Terminal entry into Khoj's local AI assistant. Use grep, get answers, never leave the shell.

“Best terminal companion for note-summarization workflows. Pipe-friendly.”

ollamallama-cppopenai-compat
Free· 4GB+ VRAM
★ 18.0k

Missing an app? Suggest one.

We curate this directory editorially — same review queue as the benchmarks feed. Open an issue with the project link and your one-line pitch for why it belongs.

Open a GitHub issue →Or contribute a benchmark →

Editorial review applies — same standards as the rest of the site. We won't list apps that don't actually work against a local runtime, regardless of marketing claims.

Where to go from here

Runtimes (/tools) →

The runtime layer: Ollama, vLLM, llama.cpp, MLX, LM Studio server, ComfyUI. What apps in this directory talk to.

Stack Builder →

Tell us your use case, get a full rig recipe — runtime + models + the apps from this directory that fit your stack.

GPU chooser →

Match an app's minimum-VRAM requirement to real hardware with our price/perf comparison.

Community benchmarks →

Real operator submissions on the model × hardware × app combos that work. The proof behind the editorial picks.