Honest by design
Cost estimates before you spend a cent. Progress that reflects real work, not spinner theater. Errors named plainly, in words.
A macOS studio for spoken word
Turn any text, into a podcast worth hearing.
Aludel turns files, links, or pasted words into a natural
conversation — on your Mac, with cloud or fully-local providers.
Read and edit the whole script before a single word is spoken.
Apple Silicon · no account · runs offline
Hear it, start to finish.
This episode was made entirely in Aludel from a chapter of a single 300-page open-access book (Free Robot Labour by Jamie Terence Kelly) — ingest, script, two-voice dialogue, mix, export. Nothing was touched by hand afterwards.
Script by Claude Sonnet, voices by Gemini Flash TTS — both via OpenRouter.
The Robot That Never Gets to Be the Boss
demo · MP3
0:00
13:02
A simple path. You stay in control.
No settings maze, no wall of options. Add your text, read and tweak the two-voice script, and you get a finished episode — Aludel takes care of everything in between.
- 1 Add your text
- 2 Review & edit the script
- 3 Get your episode
Add your content
Feed it anything you can read.
A .txt, .md, .html or PDF
file, a URL, or pasted text. Drop it in and Aludel extracts clean,
readable content — even from a 300-page book.
Long sources Experimental
Books split into chapters, automatically.
Drop a whole book and Aludel reads its outline, detects the chapters, and offers to split it into separate episodes — pick the ones you want. Still experimental: detection works best on books with a proper table of contents.
You stay in control
The script is the checkpoint.
Read and edit the full two-voice dialogue before synthesis — rewrite a line, cut a tangent, change a speaker. The cheapest place to fix quality is before the audio exists, so that's where Aludel puts you.
Your finished episode
Listen, with the transcript in sync.
The finished episode plays with a waveform and a transcript that highlights the line being spoken — click to seek. Export MP3 or WAV with proper metadata. The file is yours.
Cloud, or entirely on your machine.
Bring your own OpenAI, Anthropic, or ElevenLabs keys — or run the whole pipeline offline with a local model (Ollama or Osaurus) and Piper voices. No account, no lock-in, no per-minute meter. Aludel guides the setup either way.
Guided setup
Local models, walked through.
Aludel detects Ollama — helps you install it, start it, and pull a model — or connects to Osaurus, the native Apple-Silicon (MLX) server, and lists its models. Either way, "run it all locally, for free" is a few clicks, not a terminal session.
Your providers
Swap any model or voice.
Every LLM and voice is interchangeable. Configure them once in Services; a badge shows each voice's license so you know what's safe to publish.
Built like a tool, not a toy.
Language-decoupled
Feed it English, get Italian. The model writes directly in your output language in one pass — no translate-then-speak detour.
Private & offline
With local providers nothing leaves your Mac. No account, no telemetry, no servers of ours in the loop.
One-time, not a subscription
Free forever running locally. Pro is free while in beta — a perpetual license, not a monthly meter. See Free & Pro →
Make your first episode tonight.
Aludel is in open beta. Download it free, and unlock Pro anytime from Settings — pay what you want, including nothing, no card required.
macOS 12+ · Apple Silicon · 30 MB · v0.31.0 — What's new