xil-pipeline is an audio production pipeline that takes your markdown script and produces a fully mixed, mastered podcast episode — your choice of TTS backend, zero lock-in.
// tts backends
Pick the voice quality and cost model that fits your show. Switch backends with a single flag — no code changes.
Google Text-to-Speech via the free public API. No account, no key, no cost. Perfect for prototyping, local testing, or shows on a budget.
A high-quality open-source TTS engine that runs locally. No cloud dependency, full privacy, and expressive multi-voice output.
Industry-leading voice cloning and ultra-realistic TTS. Ideal for professional productions where voice quality is non-negotiable.
// quick start
Every step of the production pipeline has a dedicated command. Scan, parse, produce, mix, and master — all from the terminal.
// features
Built for creators and developers who want full control over their audio production workflow.
Write your script in plain markdown. Assign voices per speaker, add cues, and let the pipeline handle the rest.
Assign different TTS voices to different speakers. Host, guest, narrator — each gets their own voice profile.
Swap between gTTS, Chatterbox, and ElevenLabs with a single CLI flag. No rewiring, no code changes.
xil-daw mixes segments and xil-master normalises loudness for broadcast-ready output.
Catch script errors before production. xil-scan validates voices, segments, and config before a single byte of audio is generated.
Not a CLI person? The xil-gui Gradio dashboard gives you a full point-and-click interface for every pipeline stage.
// xil-gui
xil-gui is a Gradio-powered web dashboard that wraps the entire pipeline in a clean, browser-based UI. Upload your script, pick your backend, hit produce — no CLI required.
Free to use, open source, and ready to run. No API key required to get started.