Local LLM Apps

Tool
Category
Segment
Platform / Tool
Plan / License
Monthly Price USD
Pricing Model
Free Tier / OSS
Included Usage / Limits
Local Runtime / Model Sources
App UX / Knowledge Features
API / Extensibility
Integrations / Frameworks
Deployment / Hosting
Security / Privacy
Team / Governance
Best Fit
Main Limits / Caveats
No tagline
Local LLM AppsLow-level local LLM enginellama.cppMIT / open source$0 software; hardware/model costs separateLocal inference engine, CLI and serverRepository provides C/C++ LLM inference with GGUF support and local server toolingGGUF models, quantized open-weight models, CPU/GPU backends and many architectures supported by projectCLI tools, simple server/web UI, benchmarks, quantization and low-level runtime controlsC/C++ library, local HTTP server, OpenAI-compatible endpoints, bindings and downstream appsOllama, llamafile, KoboldCpp, LM Studio-style stacks, Python bindings and embedded appsLocal binary/library, embedded runtime or self-hosted local serverNo cloud by default; security depends how the server is bound and exposedLibrary/engine-level governance only; no app team layerDevelopers needing maximum control over local inference internalsNot a polished end-user app; model setup and flags can be technical
No tagline
Local LLM AppsLocal OpenAI-compatible stackLocalAIMIT / open source$0 software; hardware/model costs separateSelf-hosted OpenAI/Anthropic-compatible local AI stackOfficial site describes LocalAI as a free OpenAI and Anthropic alternative running locally on user hardwareLLMs, embeddings, image, audio, agents and document intelligence through modular local backendsAPI-first workflow rather than a consumer chat UI; companion LocalAGI/LocalRecall expand agent/search flowsOpenAI-compatible API, Anthropic-style compatibility, modular backends and Docker deploymentOpenAI-compatible clients, LangChain, LlamaIndex, local agents, RAG and multimodal backendsDocker, local server, workstation or on-prem hostNo cloud required; endpoint exposure, model files and plugins need operator security reviewNo SaaS governance by default; suitable for self-managed internal endpointsDevelopers replacing cloud API endpoints with a local/private stackMore backend-oriented than desktop-app UX; setup can be more technical
No tagline
Local LLM AppsDesktop local LLM workstationLM StudioProprietary desktop app$0 for personal/approved use; business terms should be checkedFree desktop app with separate terms for work/business useYes, free desktop entryDocs describe local/offline use and a localhost OpenAI-style server; model downloads and hardware are separateHugging Face/open-weight models, GGUF models, MLX on Apple Silicon and Ollama/local endpointsModel discovery, download, chat, local server, playground and parameter controlsOpenAI-like local server, structured output support, tool/MCP features and local network modeOpenAI-compatible clients, MCP servers, local apps, coding tools and custom endpointsWindows, macOS and Linux desktop; localhost or local-network servingRequests to local server stay local; cloud endpoints or remote providers follow their own termsDesktop app is individual-first; business compliance depends current LM Studio termsUsers who want a polished GUI for downloading and testing local modelsNot open source; enterprise/commercial usage terms must be verified before rollout
No tagline
Local LLM AppsAll-in-one local AI workspaceAnythingLLMMIT core / self-hosted terms$0 Desktop/self-hosted entry; hosted plans separateFree local desktop and self-hosted app; hosted/business packaging separateOfficial site promotes a free Desktop app and local/enterprise model provider supportOllama, LM Studio, LocalAI, OpenAI, Azure, AWS, Anthropic and other local or enterprise LLM providersWorkspaces, document chat, RAG, agents, MCP, plugins, desktop OS workflow and multimodal chatAPI, MCP compatibility, workspace agents, embeddings/vector stores and Docker/self-hosted configurationLocal LLM runtimes, cloud providers, vector DBs, document sources, MCP and enterprise connectorsDesktop app, Docker/self-hosted server or hosted AnythingLLMDesktop/self-hosted data can stay local; provider calls follow configured backend privacySingle-user desktop and multi-user self-hosted/hosted options; licensing terms should be reviewedPrivate document chat and agent workflows without building a full RAG stackHosted pricing and enterprise controls differ from free desktop; RAG quality depends ingestion and embeddings
No tagline
Local LLM AppsDesktop AI clientChatboxGPL-3.0 community edition$0 community software; Pro/cloud/provider costs separateDesktop client with community and commercial packagingYes, community editionOfficial repo describes a desktop client for ChatGPT, Claude and other LLMs with Ollama local model supportOllama local models, OpenAI, Azure OpenAI, Claude, Gemini and custom providersDesktop chat, local data storage, prompts, multi-provider switching and cross-platform app UXCustom providers, local Ollama connection, image generation provider support and desktop settingsOllama, OpenAI, Azure, Claude, Gemini, ChatGLM and custom API endpointsWindows, macOS, Linux desktopLocal data storage; provider calls leave device according to selected model backendCommunity app is individual-first; commercial/team controls depend paid productUsers wanting a simple desktop AI client that can point at local and cloud modelsGPL obligations for community edition; exact paid feature split can change
No tagline
Local LLM AppsPortable local LLM runnerKoboldCppAGPL-3.0 / open source$0 software; hardware/model costs separateOne-file local LLM runner and UIRepository describes easy GGML/GGUF local text generation with a KoboldAI-style UI and no complex installGGUF/GGML models, llama.cpp-derived backends, selected image/audio/TTS capabilities depending buildBuilt-in web UI, story/chat modes, sampler controls, context controls and local model loadingKobold API compatibility, OpenAI-style endpoints in recent builds, CLI flags and frontend integrationsSillyTavern, KoboldAI ecosystem, local GGUF models and llama.cpp toolingSingle local executable on desktop/server; optional LAN hostFully local when used with local models; LAN exposure must be controlledNo team governance; single-user/hobbyist workflowCreative writing and roleplay users who want a self-contained local executableAdvanced configuration can be dense; large models still require strong RAM/VRAM
No tagline
Local LLM AppsDesktop local AI studioMstyProprietary desktop app$0 Free; Aurum Annual $129/user/yearFree personal plan plus paid commercial/support licenseYes, free planPricing page lists Free forever with local and online models; professional/business use requires paid licenseOllama-backed local models, MLX models on Apple Silicon, llama.cpp models and online providersSplit chats, knowledge stacks/RAG, web search, attachments, prompts library, branching and flowchatLocal model management, custom endpoints, advanced export/search in paid tier and desktop/web studio accessOllama, MLX, llama.cpp, Azure OpenAI and online/local model providersDesktop app and Msty Studio web/desktop access depending planLocal models run on the user's machine; online providers and web search may transmit dataFree individual plan; paid license covers commercial use and more advanced featuresPower users who want a polished desktop workspace with local models and RAGProprietary; commercial usage requires paid license; pricing may change
No tagline
Local LLM AppsSingle-file local LLM packagellamafileApache-2.0 / MIT components$0 software; hardware/model costs separateSingle executable model runtime and serverMozilla page describes bundling model weights, inference engine and runtime into one executable fileGGUF/open-weight models packaged with llama.cpp/Cosmopolitan runtimeDownload-run local chat/server workflow with no separate install in supported buildsOpenAI-compatible local server mode, CLI, packaged model distribution and embedded runtimeMozilla AI tooling, llama.cpp, local apps and single-file distribution workflowsSingle local executable on desktop/server across supported OS targetsRuns locally and offline after download; file provenance and model license should be verifiedNo team governance; distribution and update governance are externalPortable demos, classrooms, offline field work and low-friction local AI experimentsModel files can be large; packaging reduces setup but not hardware requirements
No tagline
Local LLM AppsAdvanced local model web UItext-generation-webuiAGPL-3.0 / open source$0 software; hardware/model costs separateOpen-source local web UI for many model loadersRepository describes a Gradio web UI for LLMs with transformers, GPTQ, AWQ, EXL2, llama.cpp and GGUF supportTransformers, llama.cpp/GGUF, ExLlama, GPTQ, AWQ, EXL2 and other local model formats/loadersChat/notebook modes, model loader controls, extensions, character prompts and generation parametersOpenAI-compatible API extension, Gradio extensions, model loaders and community pluginsHugging Face models, llama.cpp, ExLlama, Transformers, SillyTavern and local toolingLocal Python environment, one-click installers or server-style deploymentRuns locally; extensions and remote model downloads need trust reviewNo native enterprise governance; operator controls users/networkingExperimenters who need broad loader support and fine-grained generation controlsMore setup and dependency complexity than Ollama or LM Studio
No tagline
Local LLM AppsSelf-hosted chat frameworkLobeChatMIT / open source$0 self-hosted software; hosted/provider costs separateOpen-source self-hosted UI with optional cloud/provider costsRepository describes an open-source modern AI chat framework supporting Ollama, Qwen, DeepSeek and major providersOllama, OpenAI, Claude, Gemini, Qwen, DeepSeek, OpenAI-compatible endpoints and multimodal modelsModern chat UI, knowledge base, plugins, assistants, artifacts, TTS/vision and one-click deploymentsPlugin system, server database mode, auth integrations, OpenAI-compatible providers and deployment templatesVercel, Docker, serverless, local models, cloud providers and knowledge base backendsSelf-hosted web app, Vercel/serverless or private serverSelf-hosted control; provider and storage choices determine data flowIndividual and team deployment possible; full governance depends deployment/auth setupBuilders who want a polished self-hostable AI chat product with plugin UXMore app-framework complexity than a simple desktop local runner
No tagline
Local LLM AppsPrivate desktop chatbotGPT4AllOpen source / commercial use allowed per repo$0 software; hardware and model licenses separateFree desktop app and local inference ecosystemOfficial docs describe private local desktop chat, LocalDocs and no required API calls or GPU for basic useGPT4All model ecosystem, local GGUF-style models and API/provider connectionsDesktop chat, local model browser, LocalDocs private document Q&A and settings for local inferencePython SDK, local API server, bindings and desktop integrationsNomic ecosystem, local documents, Python apps and OpenAI-compatible workflowsWindows, macOS, Linux desktop and local API serverPrivate local operation when using local models; document data remains on device unless external providers are configuredNo built-in enterprise team layer in desktop appNon-technical users who want private local document chat on everyday computersModel quality is constrained by local hardware; larger models still need significant RAM/VRAM
No tagline
Local LLM AppsOffline desktop assistantJanOpen source desktop app$0 local use; provider/model costs separateFree local desktop app plus optional external provider usageDocs state local use is always free and Jan can work offline after models are downloadedLocal models via built-in runtimes/Hugging Face plus remote providers such as OpenAI-compatible APIsDesktop ChatGPT-style interface, model hub, assistants, local API server and file/chat workflowsOpenAI-compatible local server, extensions, provider routing and model managementHugging Face models, local engines, OpenAI-compatible apps and desktop OS integrationsWindows, macOS and Linux desktopOffline operation possible; cloud model calls send data to selected providerMostly individual/local workflow; team governance depends external deployment choicesUsers who want an open-source offline ChatGPT replacement with a desktop feelLocal model performance depends hardware; optional cloud services and roadmap should be checked
No tagline
Local LLM AppsLocal model runtime and appOllamaMIT / open source$0 software; hardware and model licenses separateLocal runtime, model manager and desktop/CLI workflowOfficial project runs models locally and exposes a local API; model downloads and compute are user-providedOllama model library, GGUF-derived quantized models, Llama, Qwen, Gemma, Mistral, DeepSeek and custom ModelfilesCLI, desktop installer, model pull/run workflow and simple chat loopLocal REST API, OpenAI-compatible endpoints, Docker image, Python/JS clients and Modelfile customizationOpen WebUI, AnythingLLM, Page Assist, LangChain, LlamaIndex, Continue, Cline and many OpenAI-compatible toolsLocal desktop, local server, Docker or LAN-accessible self-hosted runtimeRuns locally by default; exposed LAN/public servers need explicit network hardeningNo SaaS governance in OSS runtime; model and endpoint access are operator-managedFastest path to local model execution for developers and hobbyistsQuality and speed depend on local hardware; model licenses vary; misconfigured servers can expose local endpoints
No tagline
Local LLM AppsSelf-hosted AI chat platformOpen WebUIOpen WebUI License / source available$0 self-hosted software; hosting and model costs separateSelf-hosted web UI and platform with optional external model costsYes, source-available self-hostingOfficial repo positions Open WebUI as a user-friendly AI interface supporting Ollama and OpenAI-compatible APIsOllama, OpenAI-compatible APIs, external providers, RAG stores and tool serversChatGPT-like UI, users, workspaces, files/RAG, tools, functions, admin controls and model switchingPipelines, functions, OpenAPI tools, REST/WebSocket APIs, OAuth/LDAP/SCIM options and Docker/Helm deploymentOllama, OpenAI APIs, Kubernetes, Docker, Helm, LDAP/OAuth/OIDC and vector/RAG backendsDocker, pip, Kubernetes/Helm, local machine or private serverSelf-hosted data control; auth, network exposure and provider routing must be configured carefullyAdmin panel, users, groups and enterprise-style auth integrationsTeams that want a self-hosted internal AI portal over local and hosted modelsLicense has branding/trademark requirements; operating it securely requires admin work
No tagline
Local LLM AppsSelf-hosted multi-provider chatLibreChatMIT / open source$0 software; model/API/hosting costs separateFree self-hosted web applicationOfficial site describes LibreChat as free, open source, self-hosted and no subscriptionOllama/local endpoints plus OpenAI, Anthropic, Google, Azure, AWS Bedrock and other providersUnified chat UI, agents, files, code interpreter-style workflows, plugins and multi-model switchingMCP support, custom endpoints, plugins, OAuth/SAML/LDAP, moderation and rate limitingDocker, MongoDB, Redis, OAuth providers, local LLM endpoints and major cloud LLM APIsSelf-hosted Docker/server deploymentSelf-hosted control; conversations still go to configured model providers unless local endpoints are usedMulti-user auth, SSO options, rate limiting and admin-oriented controlsOrganizations wanting a self-hosted ChatGPT-style portal with provider flexibilityRequires operating databases and auth; MongoDB licensing/compliance should be reviewed for commercial deployments
No tagline
Local LLM AppsLocal AI app launcherPinokioFree desktop app / source available$0 software; app/model/hardware costs separateLocal AI app launcher and localhost cloudOfficial docs describe Pinokio as a local platform to install, run and automate AI apps on a user's own machineRuns local AI apps and servers rather than one model format; apps may include Ollama, ComfyUI, Whisper and web UIsOne-click install/run, built-in browser, dependency management, app discovery and local automation scriptsJSON app scripts, local runtimes, app recipes, local web servers and agent/app control featuresPython, Node.js, Bun, Git, Conda, local AI apps and web UI stacksDesktop app hosting local web apps on the user's machineLocal-first by design; scripts can execute code and must be trusted before installNo enterprise governance by default; app/script trust is user-managedNon-specialists who want one-click local AI app installationSecurity depends heavily on script provenance; not just an LLM chat application
No tagline
Local LLM AppsPersonal AI second brainKhojAGPL-3.0 / open source$0 self-hosted software; cloud/provider costs separateSelf-hostable app plus optional cloud serviceOfficial repo describes Khoj as open-source and self-hostable, scaling from on-device personal AI to cloud-scale enterprise AILocal LLMs such as llama/qwen/mistral via local providers plus online models and Khoj cloudChat with docs/web, semantic search, custom agents, automations, newsletters and apps/pluginsAPIs, agents, browser/desktop/mobile, Obsidian/Emacs integrations and local/online model routingObsidian, Emacs, browser, desktop, WhatsApp, local LLMs, web search and document storesSelf-hosted on local machine/server or hosted Khoj cloudSelf-hosting can keep docs local; cloud app and online models change data flowSelf-hosted governance is operator-managed; enterprise/cloud governance depends planPersonal knowledge bases that need local/private model options plus searchBroader second-brain/RAG app, not a minimal local model runner
No tagline
Local LLM AppsBrowser sidebar for local modelsPage AssistOpen source browser extension$0 software; backend/provider costs separateBrowser extension and web UI for local AI modelsOfficial site describes Page Assist as an open-source browser extension with sidebar and web UI for local AI modelsOllama, Chrome AI/Gemini Nano beta, OpenAI-compatible providers such as llama.cpp, LM Studio, Llamafile and vLLMBrowser sidebar, page-aware chat, web UI, PDF/document chat and search while browsingBrowser extension APIs, local provider URLs, OpenAI-compatible endpoints and page context toolsChrome/Edge-style browsers, Ollama, LM Studio, llama.cpp, Llamafile, vLLM and local documentsBrowser extension plus local model provider on desktopPage context can be sent to configured local or remote provider; browser permissions need reviewIndividual browser workflow; no enterprise governance unless managed by browser policyUsers who want local AI assistance inside the browserDepends on a separate local model backend; browser extension permissions must be understood
No tagline
Local LLM AppsLocal computer-use assistantOpen InterpreterAGPL-3.0 / open source$0 software; provider/model/runtime costs separateLocal CLI/agent that can use local or hosted modelsOfficial repo says Open Interpreter lets LLMs run code locally and provides a ChatGPT-like terminal interfaceLocal models through Ollama, LM Studio, Jan, Llamafile and provider-agnostic model configurationTerminal chat, code execution, local file/media/data workflows and approval before running codeProvider-agnostic model picker, local model guides, shell/Python/JS execution and MCP-like local workflowOllama, LM Studio, Jan, Llamafile, OpenAI-compatible APIs, local shell and browser toolsLocal CLI and desktop/agent workflowsPowerful local code execution; user approval and sandboxing discipline are criticalNo team governance by default; local permissions and review process are user-managedTechnical users who want a local AI agent operating on their computerHigh risk if users approve unsafe code; local model quality may be insufficient for complex tasks
No tagline
Local LLM AppsLocal chat interfaceSergeApache-2.0 / open source$0 software; hardware/model costs separateDockerized local chat app over llama.cppRepository describes Serge as a web interface for chatting with Alpaca through llama.cpp, fully dockerized with an APIllama.cpp-compatible local models, originally Alpaca/LLaMA-style local modelsSimple web chat UI, local conversations and dockerized app/API stackDocker API, llama.cpp backend and local web interfaceDocker, llama.cpp and local model filesSelf-hosted Docker on local machine/serverNo API keys or cloud required when using local models; Docker/network exposure must be controlledNo team governance; hobbyist/self-hosted appUsers wanting a simple historical self-hosted llama.cpp chat UIOlder project and model assumptions; verify maintenance before relying on it
No tagline
Local LLM AppsOffline desktop/mobile chatAtomic ChatOpen source desktop/mobile app$0 softwareFree local AI chat appOfficial terms say Atomic Chat is free, open source, local, requires no account/subscription and runs models on deviceBundled/local models and Atomic local inference stack; OpenAI-compatible local server on desktop buildsPrivate offline chat, desktop/mobile app UX and simple local model operationOpenAI-compatible localhost server, GitHub source and local inference componentsDesktop apps, mobile apps, local tools and OpenAI-compatible clientsmacOS, Windows, iPhone and other app targets depending release availabilityNo account/server data for local use; uninstall removes local app data per termsIndividual-first app; no team governance capturedUsers who want a zero-account local chat client with mobile/desktop reachNewer project; model catalog, platform parity and maturity should be validated
No tagline
Local LLM AppsPrivate document assistantPrivateGPTApache-2.0 / open source$0 software; hardware/model costs separateSelf-hosted private document Q&A app/frameworkProject describes private document interaction where data does not leave the execution environmentLocal LLMs, local embeddings, document ingestion and configurable model backendsDocument ingest, private Q&A, local RAG pipeline and API/server modesPython APIs, local vector stores, embeddings, LLM backends and Docker/deployment templatesllama.cpp, Ollama-compatible/local models, embeddings models and document pipelinesLocal Python app, Docker or private serverDesigned for private local documents; first-time model downloads and configured providers need reviewNo broad team governance unless deployed behind internal auth controlsUsers who want local document Q&A without sending files to SaaSMore RAG-focused than general chat; maintaining model/storage dependencies takes work
No tagline
Local LLM AppsPower-user LLM frontendSillyTavernAGPL-3.0 / open source$0 software; backend/model/provider costs separateLocally installed LLM frontend for many backendsOfficial repo says SillyTavern is locally installed and provides no hosted service or user trackingKoboldAI/CPP, Ooba, Tabby, Ollama/OpenAI-compatible APIs, Claude, OpenRouter, NovelAI and image/TTS backendsRoleplay/story UI, character cards, lorebooks, prompt controls, extensions, Visual Novel mode and mobile-friendly layoutExtension system, backend adapters, image/TTS APIs, prompt macros and community contentKoboldCpp, text-generation-webui, Tabby, OpenAI-compatible APIs, ComfyUI, Automatic1111 and TTS toolsLocal Node.js app, Docker or local network front endNo hosted service; privacy depends selected backend and whether endpoints are remoteCommunity/hobbyist governance; no enterprise admin layerPower users building local creative writing and character chat workflowsSteep learning curve; some use cases require careful content and safety policies
No tagline
Local LLM AppsOffline ChatGPT-like appLlamaGPTMIT / open source$0 software; hardware/model costs separateSelf-hosted offline chatbotRepository describes a self-hosted offline ChatGPT-like chatbot with no data leaving the deviceLlama 2 era local models and llama.cpp-python style local inference stackSimple ChatGPT-like UI, Docker compose deployment and private local chatDocker Compose, local model files and web app/API stackUmbrel, Docker, local server and llama.cpp-style local inferenceSelf-hosted Docker or umbrelOS home serverOffline/private by design once installed; model/source downloads require trust reviewNo team governance; personal/home-server orientationHome lab users wanting a simple private chatbot applianceModel stack is older; verify maintenance and model support before new production use
No tagline
Local LLM AppsNative macOS local chatLlamaChatOpen source macOS app$0 software; hardware/model costs separateFree local macOS chat appOfficial site says LlamaChat lets users chat with LLaMA, Alpaca and GPT4All models running locally on MacLocal LLaMA, Alpaca and GPT4All-style models on macOSNative macOS chat UI focused on local model interactionGitHub source, local model loading and app-level integration with macOSmacOS, GPT4All-era local models and local model filesNative macOS desktop appLocal-only model operation when using local files; model provenance and app maintenance should be checkedNo team governance; individual desktop workflowMac users wanting a lightweight native local LLM chat clientNarrow platform scope and older model assumptions compared with Jan or LM Studio