Home / Audio / Production and Craft
Updated Jun 16, 2026 · Affirmology_BeforeBed_AudioStructure_v1.md
A craft upgrade for the Studio's before-bed structure, and a general principle about openings ("headers") that applies to every audio. This is a Studio structure change; the locked demo does not change.
We will refine this by testing real renders and talking through them. Starting notes: - A real opening, not "close your eyes / take a deep breath / welcome." Name the night, give permission to do nothing and to drift, make it safe. - Descending arc: slower, lower, softer the whole way down; taper toward near-silence; a long pause at the end, never an abrupt finish. - Breath early: two or three slow breaths, exhale longer than the inhale. - Body scan: release tension region by region, lightly tied to this person's placements. - Lead with the Moon and the feeling placements; move shadow to gift as quiet recognition; present tense, already true. - Nothing activating: no calls to action, no goals, no "tomorrow you will." Remove anything that re-alerts. - Rest coda: repeat one quiet truth, then let it go to silence. No anchor, no closing instruction. - First person, this person's own chart only, no em dashes. - Delivery: soft low soundbed with no swells; a longer length option (10 to 15 minutes); slower, softer render profile (separate from the locked demo). - Honest framing: a calm, suggestible pre-sleep window, not a medical cure for insomnia.
From how Calm/Headspace sleep stories and NSDR/yoga nidra are built: - Descending arc: voice, pace, and energy start gently and get slower, lower, and quieter the whole way down, tapering toward silence so the listener slips across the edge instead of being held at it. - A settling opening (the "header"): get comfortable, nothing to achieve, permission to stop striving and to drift. This is the on-ramp that orients before any content. - Breath awareness early: track the breath, lengthen the exhale, do not force it. - A body scan: move attention through the body region by region, releasing tension. - Calm imagery: gentle, safe, low-stakes pictures that pull attention off the day's worries. - Nothing activating: no calls to action, no tension or twists, nothing that re-alerts the brain. Sleep writers deliberately remove conflict and "do this now." - End on a long pause / taper, never an abrupt close. - Soft, low, steady soundbed; predictability and repetition so the brain learns the wind-down cue.
Those audios are generic. Ours run the same proven sleep architecture, but the words are the listener's own chart. The science is how it is delivered; the blueprint is what gets whispered. That fusion is the category we can own.
Every audio needs an intentional SETUP matched to its purpose, then the chart work. The opening is not throat-clearing, it is the context that lets the rest land. A daytime confidence audio opens with a grounded arrival into the body and the day; a sleep audio opens with a descent and permission. So "smarter context and setup" becomes a required opening beat in the script spine across the board, tuned per structure.
api/structures.py (before_bed)Replace the current before_bed directive with:
Make this a before-bed subconscious-reprogramming audio that follows the proven nighttime arc. Open with a settling header of about thirty to forty-five seconds: name that the day is ending, give permission to do nothing and to drift, make it safe. Then two or three slow heart coherence breaths with a longer exhale. Then a slow descent through the body, releasing each region, anchored lightly to this person's placements. Then lead gently with the Moon and the feeling placements, moving shadow to gift as quiet recognition, present tense and already true, never activating and with no instructions. End on a rest coda that repeats one quiet truth and tapers toward silence with a long pause. Drowsy, descending, softening tone throughout. No calls to act. Content comes only from this person's own chart.
Add a named "setup/opening" beat to the script spine that the generator must fill per structure (arrival and orientation), so every audio sets context before the four-sphere work, with the sleep version being the descent above.