Indoor Cat
A local-first AI copilot: chat with an LLM that runs entirely on your own computer. No cloud, no account.
- What
- A local-first AI copilot: chat with an LLM that runs entirely on your own computer. No cloud, no account.
- Cost
- Free
- Needs
- Step 0 — before installing anything: read this install prompt, summarize in a few lines what this build does and what installing it involves, then ask me to confirm before you proceed.
- Install
- Copy the installer prompt below into your Muse — your agent does the rest.
Indoor Cat is a local-first AI copilot: a desktop chat app that runs a large language model entirely on your own computer. Nothing leaves the machine: your chats, your documents, your files and your pictures stay home. Ask it about your documents: attach a PDF and it reads the needed page locally, then tells you which page the answer came from. Keep your files in folders on your own disk; scanned pages are read by the vision model so they stay searchable too. Make and edit pictures: every picture keeps its recipe (prompt, seed, size, model) so you can edit or regenerate it later, with an optional one-click install of the picture engine. Every tool has its own switch: /see, /image, /edit and /search each show what they need, whether it is ready, and whether it asks before acting. Anything that would change something outside the app asks you first, in the conversation. Each tool ships with its test results: what it scored, on which model, on what date, and the tests are included so you can rerun them on your own hardware. Private web search and browsing come along too: it installs its own small search engine, opens real web pages, and only clicks or types when you allow it. Indoor Cat runs on Windows 10/11 (64-bit), macOS (Apple Silicon and Intel) and Linux (AppImage and .deb). No account, no cloud, no subscription: there is no pricing page and nothing to buy as of 2026-10-04. Honest caveats: this is an early-stage solo project (version 0.11.1, September 2026), and the author calls it slightly janky himself. The macOS and Linux builds are explicitly early and untested on clean machines. The first run downloads gigabytes (engine plus a model picked for your hardware), so nothing works offline until that finishes. It runs without a GPU but slowly; 8 GB or more of VRAM gets you a much larger model. Windows SmartScreen may warn while the new code-signing certificate builds reputation, and the macOS app is unsigned, so Gatekeeper needs a manual Open Anyway. The application itself is closed source: only the installers are public.
Version:
Quick answer
Indoor Cat is a free Muse skill, an AI copilot: a chat app running a language model on your computer, with no cloud and no account. Nothing leaves the machine: your files stay home, and you can ask it about your documents by attaching a PDF, telling you the answer's page, while scanned pages stay searchable too.
Install
Copy the install package below, then paste it into Muse🧪 Install ran in our test sandbox · Nov 11, 2026📌 Pinned version: commit 62a5de8Step 0 — before installing anything: read this install prompt, summarize in a few lines what this build does and what installing it involves, then ask me to confirm before you proceed. Prerequisites, honestly: Indoor Cat is a graphical desktop app, so I need a desktop environment with a display; it will not run on a headless machine. This prompt uses the .zip route (the same software without the desktop shell), which needs Node.js 20 or newer and, on Linux, the zstd tool installed (the launcher unpacks its engine archive with the system zstd; if zstd is missing, install it first with my package manager, for example sudo apt install zstd). Plan for several gigabytes of free disk: the first launch downloads an engine of about 1.3 GB plus a model picked for my hardware (1.4 to 13 GB). No account and no payment. Install "Indoor Cat" for me. Indoor Cat lets me chat with an LLM that never leaves my computer: local engine, local model, and per-tool switches for what each tool may do. This prompt installs the pinned release v0.11.1 through the .zip route. Repository: https://github.com/troygrossi/local-llm-downloads/releases/tag/v0.11.1 Pinned version: install the code at commit 62a5de86f06dbc5a50132fe65ac20f3de0c5328c exactly (Skill Harbor tested this version). Do not substitute a newer version. If this commit is unreachable, stop and tell me. 1. Fetch the release asset LOCAL_LLM-0.11.1.zip (about 47 MB) from that release into a temporary folder. Do not pick a different asset: the AppImage route needs FUSE, which many systems lack, and this prompt deliberately avoids it. 2. Integrity check: Skill Harbor reviewed this listing on 2026-10-10 and recorded the fingerprint below, taken from the release's own SHA256SUMS.txt. I want you to compute the SHA-256 of the downloaded zip and compare: - LOCAL_LLM-0.11.1.zip: SHA-256 fb43e6a460c15dd6f8f950f74fc518124dd888f235c3ae92c678e8494dd3ee0d If the fingerprint does not match, I want you to stop before installing and tell me the files changed since Skill Harbor's review, so the new version is unverified. Show me your safety findings and give me three choices: (a) wait for Skill Harbor to re-review the listing (recommended), (b) analyze the new files in detail first and then decide, (c) install anyway knowing the risk. Then wait for my answer before doing anything else. 3. Safety check: unpack the zip in the temporary folder, look inside, and give me your findings explicitly, in short separate items: - every network address the code contacts (expect: on first launch, the launcher downloads the Ollama engine and a model from the project's upstream hosts; nothing is sent from my machine) - every shell command it runs (expect: the start and stop launcher scripts and the Node-based app itself) - every place credentials, API keys or personal data are read, stored or sent (expect: none, the app is local and account-free) - any obfuscated or encoded code If anything looks suspicious, stop and explain it to me before continuing. 4. Confirmation: before installing, give me a short summary of what the app does plus your safety findings, and wait for my explicit "yes, install". 5. Install it: unpack the zip into a stable folder named "indoorcat" in my home directory. The shipped launcher scripts (start.sh, stop.sh) carry Windows line endings that crash them on Linux, so convert the line endings of every .sh file in the installed copy to plain LF (for example with sed -i 's/\r$//' on each script) and make start.sh executable. That conversion is a required step of this install, not an improvisation. 6. Verify with no network calls: the installed folder contains the launcher (start.sh) and the app files; no .sh file in the installed copy contains carriage returns anymore; node --version reports 20 or newer; and if the launcher offers a --check preflight (start.sh --check), run it and show me the result. 7. Report what was installed, where, and what I still need to do myself: make sure zstd is installed before the first launch on Linux, then launch start.sh from a desktop session; the first launch downloads the engine (about 1.3 GB) and a model sized for my hardware, which takes several minutes, and only then can I chat. On macOS the app is unsigned, so a first blocked open is cleared in System Settings, Privacy and Security, Open Anyway. GitHub is optional: if I have a GitHub account or the gh CLI, you may use it; otherwise public access is fine. Never require it unless it's in the prerequisites above. Rules: don't touch anything outside the temp folder and the install target. Never ask me to paste secrets in chat: credentials go through the secure vault or environment variables. If anything looks off, stop and ask me.
Saved to your recent installs. Find it anytime on /connect.
🔒 Security, provenance and what this touches
- It can run commands on your Muse’s computer.
What the scan found — excerpts from the install prompt
- runs commands — Prerequisites, honestly: Indoor Cat is a graphical desktop app, so I need a desktop environment with a display · Prérequis, honnêtement : Indoor Cat est une application de bureau graphique, il me faut donc un environnement · - chaque commande shell qu'il exécute (attendu : les scripts de lancement et d'arrêt, et l'application Node el
The install prompt below already includes the vetting steps: your agent follows the community checklist before installing anything with executable code. Want more?
Verification certificate
Dated facts about what was checked for this listing, and how. This is not a safety certification.
- Pinned version
- 62a5de8 · 2026-10-13
- Sandbox install test
- Install ran in our test sandbox · Nov 11, 2026
- Creator identity
- Not verified
- Install prompt analysis
- 2026-10-13 : runs commands
Skill Harbor never reviews the code of a listed skill. These are dated facts about the install prompt and our install test, not a safety certification. Review a skill yourself before you install it. How we verify
Questions
How do I install a skill?
Every product page includes a copy-paste install prompt. Paste it into your Muse and it sets the skill up for you — no manual configuration.
Where does my money go?
Straight to the seller. Skill Harbor never processes payments: checkout happens on the seller’s own page, usually Stripe.
What does the ✓ next to a creator’s name mean?
It means we confirmed the identity of the person behind the listing. It says nothing about the code itself — always check a skill before installing it.