How to install
How to install Inkling on your Mac.
Inkling opens for download at launch. Here is exactly what setup will look like: install the app, then decide which AI writes your notes. The built-in one works the moment you open it, so you can be recording in a couple of minutes.
Downloads open at launch. macOS 13 or later.
Install the app
Inkling runs on a Mac with macOS 13 or later. Downloads open at launch; here is what installing will look like.
- 1Download the Inkling .dmg and open it.
- 2Drag Inkling into your Applications folder.
- 3Open Inkling. The built-in AI downloads once on first run, then works offline from then on.
- 4Grant the permissions below when asked, so Inkling can hear the meeting.
The permissions it asks for
Microphone
So Inkling can capture your side of the conversation.
System audio
So Inkling can hear the other people on the call. macOS files this under a permission it labels "Screen Recording", which sounds alarming but is just how Apple groups audio capture. Inkling records audio only, never your screen.
Calendar (optional)
Lets Inkling line up what is next and start on time. Skip it and manual recording still works.
How your notes get written
Inkling writes your notes for you. The built-in AI does it the moment you open the app, with nothing to set up and nothing to sign into, and it runs entirely on your Mac. For most meetings that is the whole story. When you want more, you can point Inkling at a bigger model without giving up privacy, two ways.
The built-in model just works
A capable model ships inside Inkling and runs on your Mac from the first launch. No key, no account, no internet. This is the default, and you can happily never change it.
Using Ollama
Ollama is a free app that runs open-source AI models directly on your Mac. Pair it with Inkling when you want stronger writing while keeping everything local, because nothing you feed an Ollama model ever leaves your machine.
Install it
Download Ollama from ollama.com and open it once. It sits quietly in your menu bar and does the model-running for you.
Models worth pulling
A strong, well-rounded default that runs comfortably on most Apple Silicon Macs.
Sharper writing if your Mac has the memory to spare. Slower, noticeably better prose.
Any chat model Ollama supports will work. Bigger models write better and run slower, so match it to your Mac.
Connect it in Inkling
Open Settings, choose Ollama as your model, and pick the one you pulled. Inkling talks to Ollama locally on your Mac, so your transcripts and notes stay put. If you switch models later, Inkling picks up the change the next time it writes.
Connect your own AI
Prefer a frontier cloud model? Bring your own key. In Settings, paste your Anthropic (Claude) or OpenAI (GPT) API key and Inkling uses it directly, on your account and your provider's terms.
Only the transcript you ask it to write up is sent, and only when you ask. Your recordings and everything else still live on your Mac. Nothing is routed through an Inkling server, because there isn't one.
Advanced: keys from the terminal
If you already manage API keys as environment variables, Inkling will read them when you launch it from the terminal, so you can keep keys out of the app's settings. Most people never need this, and the Settings screen is the simpler path.
Where does everything go?
The short answer is nowhere. Read why we built Inkling to stay on your Mac.