jc.JACOB CLARKENGINEER. BUILDER. GAME MAKER.

GOOD THINGS START WITH A CONVERSATION

Let’s make
something interesting.

A tricky problem, a new idea, or just a hello. Here’s where to find me.

Copy the address, or choose a way to say hello below.

Open email app Join the Discord
← AI LAB / DESKTOP TOOLS

Markdown Podcast Studio LATEST RELEASE · MAC + WINDOWS

Turn your notes into a conversation worth listening to.

Drop in Markdown or research notes. Draft an editable two-host podcast, choose distinct local voices, and make an MP3 you can listen to anywhere.

Choose your download ↓How to get started
Episode workspace with a synthetic two-host script, source intake, and local voice choices.
Episode workspace with a synthetic two-host script, source intake, and local voice choices. Synthetic demo data in the real desktop app.

01 / DOWNLOAD

Your studio. On your computer.

Desktop release 0.2.1. Extract the full ZIP before opening the app.

Mac · Apple Silicon

M1 or newer · macOS 12+

Download ZIP ↓ 212 MB
SHA-256 checksumda8075cba4e12c2a192d722ca3910e4dd2402bfcf10389625b2bfb551b69f939

Mac · Intel

Intel processor · macOS 12+

Download ZIP ↓ 218 MB
SHA-256 checksumb925a67739e078f2250ef0019ace5233941f72caccc19b4263bef9980b120a54

Windows · 64-bit

Windows 10/11 · Intel or AMD x64

Download ZIP ↓ 261 MB
SHA-256 checksumf0cc34f77b36fd49215eb0f8b664fbcc9c1acfd0ee4b3dd9d79c3205703b553b

These early builds are unsigned and not notarized. macOS or Windows may warn or block them because there is no verified publisher signature. No security-bypass launcher is included.

Download all checksums

Give ideas two voices.

Draft a talk-show style exchange with an optional writing provider. Edit every line and choose a distinct Kokoro voice for each host.

Keep the script yours.

Paste notes or drop Markdown/text files. Optional codebase, Wikipedia, and web-source research can inform a draft; the final script stays editable.

Take the episode with you.

Render voice audio locally and export MP3 through FFmpeg. Single-narrator mode reads prose directly without an AI writer.

02 / FIRST SESSION

Notes in. An episode out.

Electron includes the desktop runtime. MP3 export needs FFmpeg installed separately and on PATH. The Kokoro voice model downloads on first use. Automatic dialogue drafting needs LM Studio or a separately installed, signed-in Codex/Claude CLI.

  1. Extract and open.

    Mac: unzip and move MarkdownPodcastNarrator.app to Applications. Windows: extract the entire folder and open MarkdownPodcastNarrator.exe; keep the companion files together.

  2. Prepare the tools you need.

    Install FFmpeg using its official download guide at ffmpeg.org/download.html. For automatic drafting, start an LM Studio writing model locally or select your own signed-in CLI provider. No account or model is bundled.

  3. Build the conversation.

    Drop a Markdown/text file, choose two voices, select your writer, and click Draft. Edit the Host A/Host B script. Or turn off two-host conversation to read prose directly.

  4. Present and keep the MP3.

    Click Present & play. The first voice render downloads the Kokoro model. Finished files save in Music / Markdown Podcast Narrator; use Show file to find the result.

03 / RELEASE NOTES

A working desktop release, with clear limits.

Native CI checks startup, the rendered interface, and Kokoro/ONNX imports on each platform. These checks do not validate a complete generated episode, audio quality, or every writing provider. Notes and research sent to a CLI writer use your provider account and may consume limits or incur charges. The display name has changed; the project ID and local data location are preserved.

Documentation images capture only the app’s own interface with synthetic fixtures. No screen, camera, or microphone recording, personal media, voice cloning, or paid AI generation was used.