{
  "listings": [
    {
      "slug": "llama-cpp",
      "name": "llama.cpp",
      "company": "ggml.ai, part of Hugging Face since 2026",
      "grade": "D",
      "fit": "yes",
      "why": "--offline forces the model cache and prevents network access; no telemetry is documented or shipped, and --api-key is available. An account serves Claude Code from it on local hardware.",
      "source": "https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/llama-cpp"
    },
    {
      "slug": "anythingllm",
      "name": "AnythingLLM",
      "company": "Mintplex Labs, Inc.",
      "grade": "E",
      "fit": "caveat",
      "why": "Telemetry (PostHog) is on by default with DISABLE_TELEMETRY or an in-app switch. The README lists the outbound connections that remain, such as the model CDN. Local LLM and embedder options are documented.",
      "source": "https://github.com/Mintplex-Labs/anything-llm#telemetry--privacy",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/anythingllm"
    },
    {
      "slug": "berth",
      "name": "Berth",
      "company": "Trevin Tindall (T-Crypt on GitHub)",
      "grade": "B",
      "fit": "caveat",
      "why": "A loopback daemon and gateway with no telemetry mentioned; berth update and the docs mirror reach the network. No privacy page.",
      "source": "https://github.com/T-Crypt/berth/blob/main/README.md",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/berth"
    },
    {
      "slug": "bitrun",
      "name": "BitRun",
      "company": "Abhishek Khanra (IshekKhal on GitHub)",
      "grade": "E",
      "fit": "caveat",
      "why": "Models download from Hugging Face and offline archives are on the releases page; the energy telemetry is local power measurement. No privacy statement, no checksum on downloads.",
      "source": "https://github.com/IshekKhal/bitrun/blob/main/README.md",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/bitrun"
    },
    {
      "slug": "foundry-local",
      "name": "Foundry Local",
      "company": "Microsoft Corporation",
      "grade": "B",
      "fit": "caveat",
      "why": "The overview says data never leaves the device and the app works offline, listing model and execution-provider downloads as the network uses. The telemetry opt-out is not on the Learn pages read.",
      "source": "https://learn.microsoft.com/en-us/azure/foundry-local/what-is-foundry-local",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/foundry-local"
    },
    {
      "slug": "ghost-core",
      "name": "Core",
      "company": "ZMJ, Inc., doing business as Ghost, 325 9th Street, San Francisco, CA 94103, United States",
      "grade": "F",
      "fit": "caveat",
      "why": "The FAQ and setup guide say inference and memory stay on the device and prompts are not sent to a cloud, while web search, email and updates need the internet. Nothing can be verified until units ship.",
      "source": "https://ghost.ai/core",
      "confidence": "low",
      "url": "https://www.anchorterminal.com/tools/ghost-core"
    },
    {
      "slug": "gpt4all",
      "name": "GPT4All",
      "company": "Nomic, Inc.",
      "grade": "F",
      "fit": "caveat",
      "why": "Runs offline with local models and accounts use it to keep chats off third-party servers. No telemetry or data-flow page was found, and the last release is February 2025.",
      "source": "https://news.ycombinator.com/item?id=45817867",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/gpt4all"
    },
    {
      "slug": "hyperqwen",
      "name": "HyperQwen",
      "company": "syv.ai (Copenhagen, CVR DK44671247)",
      "grade": "C",
      "fit": "caveat",
      "why": "Runs on the owner's card with weights pulled from Hugging Face on first start and an image from the registry; usage statistics are disabled in the Dockerfile per the earlier reading. No privacy page.",
      "source": "https://github.com/syv-ai/HyperQwen/blob/main/README.md",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/hyperqwen"
    },
    {
      "slug": "jan",
      "name": "Jan",
      "company": "Menlo Research Pte Ltd",
      "grade": "C",
      "fit": "caveat",
      "why": "The server page describes private, offline use and a Request Timeout for slow hardware. No data-flow or telemetry page was found, and Hugging Face downloads route through Menlo's mirror.",
      "source": "https://github.com/janhq/jan/blob/main/docs/src/pages/docs/desktop/api-server.mdx",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/jan"
    },
    {
      "slug": "khoj",
      "name": "Khoj",
      "company": "Khoj Inc.",
      "grade": "E",
      "fit": "caveat",
      "why": "Offline chat with local models is supported. Anonymised telemetry goes to PostHog by default with a documented opt-out. Clients default to app.khoj.dev, which has closed.",
      "source": "https://docs.khoj.dev/privacy",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/khoj"
    },
    {
      "slug": "koboldcpp",
      "name": "KoboldCpp",
      "company": "LostRuins (Concedo)",
      "grade": "B",
      "fit": "caveat",
      "why": "A single executable serving GGUF files locally with --password and --host. No privacy or telemetry statement is published either way.",
      "source": "https://github.com/LostRuins/koboldcpp/blob/concedo/README.md",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/koboldcpp"
    },
    {
      "slug": "lemonade",
      "name": "Lemonade",
      "company": "Advanced Micro Devices, Inc.",
      "grade": "B",
      "fit": "caveat",
      "why": "Models run locally and Cloud Offload is opt-in and experimental. No privacy, telemetry or data-flow page was found in the docs read.",
      "source": "https://github.com/lemonade-sdk/lemonade/blob/main/README.md",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/lemonade"
    },
    {
      "slug": "lm-studio",
      "name": "LM Studio",
      "company": "Element Labs, Inc.",
      "grade": "C",
      "fit": "caveat",
      "why": "The offline page says chat, RAG and the server work without internet and lists what needs connectivity (search, downloads, runtimes, update checks). No telemetry statement or switch on that page. One account keeps it as a fallback.",
      "source": "https://lmstudio.ai/docs/app/offline",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/lm-studio"
    },
    {
      "slug": "localai",
      "name": "LocalAI",
      "company": "Ettore Di Giacinto and the LocalAI team",
      "grade": "B",
      "fit": "caveat",
      "why": "Everything runs in local backends pulled on demand; no telemetry or data-flow page was found in the docs read.",
      "source": "https://github.com/mudler/LocalAI/blob/master/README.md",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/localai"
    },
    {
      "slug": "mlx-lm",
      "name": "MLX LM",
      "company": "Apple Inc.",
      "grade": "C",
      "fit": "caveat",
      "why": "Models load from the local Hugging Face cache after the first download and no telemetry is mentioned, but nothing documents offline behaviour or an outbound inventory.",
      "source": "https://github.com/ml-explore/mlx-lm/blob/main/mlx_lm/SERVER.md",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/mlx-lm"
    },
    {
      "slug": "ollama",
      "name": "Ollama",
      "company": "Ollama Inc.",
      "grade": "D",
      "fit": "caveat",
      "why": "One account ran Claude Code offline on a plane against Ollama. The docs list no inventory of outbound connections or a telemetry switch, and cloud-tagged models run on Ollama's servers.",
      "source": "https://www.williamangel.net/blog/2026/04/27/offline-agentic-coding.html",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/ollama"
    },
    {
      "slug": "open-webui",
      "name": "Open WebUI",
      "company": "Open WebUI Inc.",
      "grade": "D",
      "fit": "caveat",
      "why": "OFFLINE_MODE and HF_HUB_OFFLINE stop downloads and update checks, but the page says they do not block outbound connections and tells you to use firewall rules; analytics preferences exist.",
      "source": "https://docs.openwebui.com/tutorials/maintenance/offline-mode",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/open-webui"
    },
    {
      "slug": "privategpt",
      "name": "PrivateGPT",
      "company": "PriBAI Technology Spain S.L.",
      "grade": "C",
      "fit": "caveat",
      "why": "Everything runs against a local model server and tracing is written to local OTLP files. Web search and fetch tools are opt-in; no telemetry statement was found.",
      "source": "https://docs.privategpt.dev/introduction",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/privategpt"
    },
    {
      "slug": "screenpipe",
      "name": "screenpipe",
      "company": "Negentropy Labs, Inc. (dba Screenpipe)",
      "grade": "C",
      "fit": "caveat",
      "why": "The data-flow page tabulates what leaves the machine (cloud transcription, cloud AI, sync, media analysis) with settings to turn each off, and local Whisper is an option. Analytics are not in that table.",
      "source": "https://docs.screenpipe.com/privacy-data-flow",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/screenpipe"
    },
    {
      "slug": "text-generation-webui",
      "name": "TextGen",
      "company": "oobabooga",
      "grade": "E",
      "fit": "caveat",
      "why": "Portable builds run locally with an optional key; no telemetry or privacy statement exists either way.",
      "source": "https://github.com/oobabooga/textgen/blob/main/README.md",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/text-generation-webui"
    },
    {
      "slug": "vllm",
      "name": "vLLM",
      "company": "The Linux Foundation (vLLM is a PyTorch Foundation project)",
      "grade": "D",
      "fit": "caveat",
      "why": "Anonymous usage statistics go to stats.vllm.ai by default; VLLM_NO_USAGE_STATS, DO_NOT_TRACK or a do_not_track file switches them off.",
      "source": "https://docs.vllm.ai/en/stable/usage/usage_stats/",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/vllm"
    },
    {
      "slug": "docker-model-runner",
      "name": "Docker Model Runner",
      "company": "Docker, Inc.",
      "grade": "C",
      "fit": "unchecked",
      "why": "docs.docker.com is paused for automated reading under Docker's terms, and the README says nothing about telemetry or offline use.",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/docker-model-runner"
    },
    {
      "slug": "underdog",
      "name": "Underdog",
      "company": "Conway Research",
      "grade": "F",
      "fit": "unchecked",
      "why": "Conway says everything stays on the machine, but underdog.ai refuses our reader and the card has no data-flow statement.",
      "confidence": "low",
      "url": "https://www.anchorterminal.com/tools/underdog"
    },
    {
      "slug": "localghost",
      "name": "LocalGhost",
      "company": "LocalGhost.ai Ltd (company number 17213100, registered in England and Wales, incorporated 12 May 2026)",
      "grade": "F",
      "fit": "no",
      "why": "The README lists every outbound request and calls any other a bug, which is exemplary, but nothing serves an agent or runner, so the use case does not apply.",
      "source": "https://github.com/LocalGhostDao/localghost/blob/main/README.md",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/localghost"
    }
  ],
  "marks": "yes: the docs show it does the job as the success test says. caveat: it does the job with a limit an agent should know. no: it doesn't do this job. unchecked: we couldn't read it.",
  "meta": {
    "attribution": "Anchor Terminal (https://www.anchorterminal.com)",
    "attributionRequired": true,
    "credit": "CC BY 4.0, credit Anchor Terminal and link the page or API address you took it from (https://www.anchorterminal.com/terms/#data)",
    "docs": "https://www.anchorterminal.com/docs/",
    "generatedAt": "2026-11-16",
    "license": "CC-BY-4.0",
    "method": "https://www.anchorterminal.com/benchmark/",
    "methodology": "0.8",
    "openapi": "https://www.anchorterminal.com/openapi.json",
    "preview": false,
    "run": "2026-10-01",
    "runLabel": "October 2026 research run",
    "terms": "https://www.anchorterminal.com/terms/#data"
  },
  "useCase": {
    "accounts": 6,
    "after": [
      "local-ai.coding-agent-backend"
    ],
    "area": "models",
    "asks": [
      "I'm on a plane, keep coding",
      "the cloud provider is erroring, switch to local",
      "I don't want my chats on anyone's server",
      "sandbox the agent and the model together"
    ],
    "before": [
      "inference.failover-routing"
    ],
    "blockers": [
      "telemetry on by default with the opt-out covering only part of it",
      "a model catalogue fetched over plain HTTP, so the first download is the exposure",
      "a feature (live transcription, web search) that is a cloud call and fails with a 503 when the vendor's service is down"
    ],
    "capabilities": [
      "inference.local",
      "inference.open-weights"
    ],
    "category": "local-ai",
    "categoryName": "Local AI",
    "checked": "2026-11-12",
    "companies": {
      "fit": 1,
      "fitWithLimit": 20,
      "tagged": 49
    },
    "competition": "one",
    "fits": [
      "llama-cpp"
    ],
    "fitsWithLimit": [
      "anythingllm",
      "berth",
      "bitrun",
      "foundry-local",
      "ghost-core",
      "gpt4all",
      "hyperqwen",
      "jan",
      "khoj",
      "koboldcpp",
      "lemonade",
      "lm-studio",
      "localai",
      "mlx-lm",
      "ollama",
      "open-webui",
      "privategpt",
      "screenpipe",
      "text-generation-webui",
      "vllm"
    ],
    "id": "local-ai.private-offline-run",
    "listings": {
      "caveat": 20,
      "no": 1,
      "notRead": 0,
      "tagged": 50,
      "unchecked": 2,
      "yes": 1
    },
    "name": "Run with nothing leaving the machine, offline or as a fallback when the cloud fails",
    "status": "draft",
    "success": "The runner serves requests with the network interface down, and telemetry has a documented off switch. Measured: a network monitor during a run and during indexing shows no outbound connections other than those the owner enabled.",
    "url": "https://www.anchorterminal.com/api/v1/use-cases/local-ai.private-offline-run.json",
    "weight": 7,
    "weightBasis": "6 of 50 accounts name privacy, offline use or cloud failover as the reason; the buyer check on network traffic sits here"
  }
}
