SMF·ALITU
Can a prompt replace Alitu?
Audio & video editing — podcast recording, cleanup and publishing
Exhibit tracking slip
Verdict
Trimming silence, normalizing loudness, and stitching an intro, main recording, and outro into one exportable episode is a real, working loop — but getting the audio-processing chain (noise gating, level matching across segments, fade timing) genuinely reliable across different recordings takes a weekend of tuning, not a sitting. What's missing entirely is Alitu's hosting and directory distribution.
Exhibit A — The prompt
Received on31.07.2026Build a local podcast episode assembler: upload raw audio segments, clean them up automatically, and export a finished, ready-to-upload episode file. Stack: Python, FastAPI, ffmpeg for all audio processing, SQLite for episode metadata. No cloud audio API — all processing runs locally through ffmpeg filters.
Core loop: the user uploads a main recording plus optional intro and outro clips, keeping originals untouched in a read-only folder. Processing runs silence trimming, loudness normalization to -16 LUFS (the common podcast target), a light noise gate, and crossfades between segments, writing results as a new file rather than overwriting the source. A simple two-track timeline view lets the user reorder segments and preview before export. On export, generate a local transcript (via whisper.cpp) alongside the final MP3 and WAV, plus a JSON manifest recording every processing parameter used, so a run can be reproduced or adjusted.
This needs no external API key — whisper.cpp runs the transcript locally, and all audio processing is ffmpeg. If skipped, transcription is simply blank and the episode still exports fine.
Do not build: podcast hosting, RSS feed generation, or directory submission — those require a public server and ongoing bandwidth this tool doesn't provide. Do not build a licensed music library; the user supplies their own intro/outro clips. Do not build remote multi-track recording — this assembles audio that already exists, it doesn't capture it.Opening prefills the prompt — press enter to run it.
Exhibit B — What you lose
- B.1 one-click podcast hosting and RSS feed generation
- B.2 distribution to Spotify, Apple Podcasts, and other directories
- B.3 a licensed royalty-free music library for intros and outros
- B.4 advanced mastering tuned per-voice rather than one fixed normalization target
- B.5 support when an export doesn't sound right and you don't know why
Prior art
Exhibit C — Why people still pay: audio infrastructure, distribution, and production polish
People pay for Alitu because cleaning up one episode is buildable, but doing it reliably across dozens of episodes with different microphones, rooms, and voices — without ever second-guessing whether this week's export actually sounds right — is the part that takes real audio engineering, not just an ffmpeg command.
Questions
Can I import my existing Alitu projects?
No — Alitu's project files and processing settings are internal to their app. You'd start with your raw recordings and re-create your normalization and fade preferences once, then reuse them for future episodes.
Will it work on my phone?
No — audio processing through ffmpeg needs real compute and file access this build assumes a desktop or server has. You'd record on your phone if you want, then process on a computer.
What does it cost to run?
Nothing per episode — ffmpeg and whisper.cpp both run locally with no per-minute charge. The only cost is wherever you host the app itself, or nothing if it just runs on your own machine.
What's the one thing that doesn't survive the rebuild?
Distribution. Alitu's real endpoint is a live RSS feed that Spotify and Apple Podcasts pull from automatically; this build hands you a finished MP3 file, and getting it in front of listeners is still a separate hosting step you'd have to solve yourself.
Related tools
Receipt