Medita

A Quiet Markdown Editor for macOS. A sharper reader for your ideas

A quiet editor for Markdown. A sharper reader for your ideas.

Medita's AI companion runs DeepSeek R1 entirely on your Mac — no account, no subscription, no data sent anywhere. Or connect your own API key for faster cloud analysis. Your writing stays yours either way.

Medita Logo

The Vision Behind Medita

Most professional writing begins and ends in Markdown. Whether it is drafting academic papers, documenting complex synthesis patches, or preparing tutorials for Moodle, Markdown provides the necessary structure without the bloat.

However, the creative process often feels fragmented. Writers are frequently forced to cycle through multiple tools: one for drafting, another for media management, and yet another for final export. Medita was designed to bridge these gaps. It is a dedicated environment for composers, sound artists, and technical writers who require a tool that handles audio and images as first-class citizens while maintaining a distraction-free interface.

What it does

Medita keeps the writing experience minimal while covering everything a serious writer needs.

Write in Markdown with ATX (# Heading) or Setext (Heading\n===) heading styles, switchable at any time. Full support for bold, italic, underline, blockquotes, lists, code blocks, horizontal rules, and links.

Embed images and audio using a clean syntax inspired by Pandoc. Drop any image or audio file directly onto the editor and it inserts the correct snippet automatically:


@[Dawn chorus, recorded at 5:14 am](https://www.google.com/search?q=/path/to/audio.wav)

Audio embeds render as a native <audio> player on export — useful for sound designers, musicians, and anyone writing documents that live alongside media.

Export to HTML or DOCX. The HTML export produces a clean, self-contained document with all styles inlined — no external dependencies. The DOCX export uses proper heading styles, so your document arrives in Word already formatted.

An AI writing companion built for serious work. Hit Review and Medita reads your draft the way a careful reader would — identifying logical inconsistencies and unsupported claims, structural gaps, and arguments that work against themselves.

Medita runs DeepSeek R1 locally on your machine using llama.cpp — no account, no subscription, no data leaving your computer. On an M-series Mac, expect generation times between 20 and a 40 seconds. If you want almost instant results, connect your own API key from Groq, Together, OpenRouter, or DeepSeek's own API under AI Engine settings — same model, same analysis, text sent to your chosen provider.1

Example export

Here is a sample document written in Medita and exported to HTML — a short field recording journal entry using headings, bold and italic text, a blockquote, an image, and an audio embed.

# Field Notes — March

## Dawn chorus, old-growth forest

I arrived before sunrise, while the fog was still **dense enough to muffle footsteps**. The forest sounded different — closer, more contained. By 5:14 am the birds had started, layered and overlapping in a way that no single recording can fully resolve.

> The acoustics of an old-growth forest are unlike anything a room can reproduce. The canopy acts as a diffuser; the ground, wet with dew, as an absorber. What reaches the microphone is already a mix.

I set up two cardioid microphones in an *ORTF configuration*, about two metres off the ground, pointing into the canopy. The recording ran for forty minutes. The excerpt below captures the first wave of activity just after first light.

![Morning mist filtering through the canopy, old-growth forest](/images/forest.png)

@[Dawn chorus recorded at 5:14 am — ORTF, 40m excerpt](/audio/birds.mp3)

Field Notes — March


Dawn chorus, old-growth forest


I arrived before sunrise, while the fog was still dense enough to muffle footsteps. The forest sounded different — closer, more contained. By 5:14 am the birds had started, layered and overlapping in a way that no single recording can fully resolve.


The acoustics of an old-growth forest are unlike anything a room can reproduce. The canopy acts as a diffuser; the ground, wet with dew, as an absorber. What reaches the microphone is already a mix.


I set up an Ambisonics microphone, about two metres off the ground, to capture the sound field. The recording ran for forty minutes. The excerpt below captures the first wave of activity just after first light.


Morning mist filtering through the canopy, old-growth forestMorning mist filtering through the canopy, old-growth forest

Dawn chorus recorded at 5:14 am — Ambisonic, 40m excerpt

Features at a glance

System requirements

License

Medita is sold as a one-time purchase. Your license covers installation on up to two personal devices. No subscription, no expiry.

Buy a Medita License — $29

Questions? Contact us: info@wandersound.ca

1

Performance metrics on M-series hardware using Medita Timing: 101 words (1,604 prompt tokens) at 17.8 t/s (9.28s encode, 27.53s gen); 438 words (2,010 prompt tokens) at 16.8 t/s (9.27s encode, 44.34s gen).