Erm
title: erm type: tool created: 2026-07-13 last_updated: 2026-07-20 related: [] sources: ["https://github.com/dougcalobrisi/erm"] radar_quadrant: Tools radar_ring: retired radar_position: outer
erm is a local command-line tool that removes disfluencies — filler sounds such as "um," "uh," "er," "ah," "hmm," and "uh-huh" — from recordings of English speech. It runs entirely on the user's machine against an audio file and produces a cleaned recording with those sounds cut out.
The tool targets anyone editing spoken-word audio — podcast producers, voice-memo transcribers, video creators — who would otherwise remove filler words manually in an audio editor. It is maintained on GitHub under an MIT license, has drawn 110 stars, and was last pushed 2026-06-14 with 2 open issues.
erm is retired from the Tools quadrant as disqualified. It entered at Assess/outer on 2026-07-13 as a narrow single-purpose utility. A later first-person evaluation ran the tool against real recordings and found the output poor quality: it did not deliver usable disfluency removal. The disqualified retirement records that verdict so the tool is not revisited. The outer position is auto-derived from the disqualified reason and marks it as known-bad rather than merely niche.
