Stories

Full story runtime + prompt + provider configuration, directly backed by AdminCore callables.

Prompt chain editor

Story Browser

Story Mode Access Status
No stories match this filter.

Story Voice Settings

Optional. This is the primary preview text in iOS story browsing.
When enabled, every new session starts with this authored welcome beat before dynamic story generation begins.
Optional short label used as the beat name.
This is hard-authored copy for the first beat of every new session.
Canonical runtime engine for all story sessions.
Enable map element retrieval and passbook support.
Enable next-beat generation for this story.
Required when beats are enabled.
Enable user-created/editable elements for supported browse experiences (auth required).
Region scope editing requires preconfigured regions; use CLI for region geometry updates.
Strict contract: object with dataLayers[], wikipediaEnabled, and googlePlacesEnabled.
Member-only stories use the membership/paywall flow. Free stories bypass membership gating.
Published stories appear in the iOS app. Unpublished stories are drafts.

Story Cover Art

Generate square abstract cover candidates for the selected story, then approve one to set it live.

Generate candidates to review and promote a story cover.
1

Place Selection

Settings that affect how the AI picks the next place

Used as the full system prompt. Edit freely to change scoring, format, and JSON response requirements.
Placeholders: {{userLat}}, {{userLng}}, {{candidatePlaces}}, {{narrativeHistory}}, {{locationHistory}}
Place Picker Line Templates (Advanced)
Placeholders: {{index}}, {{title}}, {{distanceMeters}}, {{lat}}, {{lng}}
Placeholders: {{index}}, {{placeName}}, {{themes}}, {{narrative}}
Placeholders: {{placeName}}
Provider used for place selection.
Model used for place selection decisions
Selection variance: 0=deterministic, 2=creative.
How far to look for Wikipedia places (default: 500m)
Max places to consider for scoring (default: 10)
"RIGHT HERE" - full visual details (default: 30m)
"SHORT WALK" - moderate details (default: 100m)
"FURTHER AWAY" - no visuals (default: 500m)
2

Narrative Generation

Settings that affect how the story is written

Placeholders: {{wordCountGuidance}}, {{distanceGuidance}}, {{scaleGuidance}}, {{angleGuidance}}, {{languageInstruction}}, {{wordCountMin}}, {{wordCountMax}}
Placeholders: {{placeName}}, {{contextSection}}, {{locationSection}}, {{continuityPrompt}}, {{transitionPrompt}}, {{priorNarratives}}, {{locationHistory}}, {{previousThemes}}
Narrative Templates (Advanced)
Placeholders: {{placeName}}, {{contextSection}}, {{locationSection}}
Placeholders: {{context}}
Placeholders: {{lat}}, {{lng}}
Placeholders: {{index}}, {{placeName}}, {{wikipediaUrl}}, {{text}}
Placeholders: {{placeName}}, {{wikipediaUrl}}
Placeholders: {{themes}}
Narrative Guidance Templates (Advanced)
Placeholders: {{wordCountMin}}, {{wordCountMax}}
Placeholders: {{thresholdClose}}
Placeholders: {{thresholdClose}}, {{thresholdNearMax}}
Placeholders: {{thresholdNearMax}}, {{thresholdFar}}
Placeholders: {{thresholdFar}}
Placeholders: {{suggestedAngle}}
Placeholders: {{languageName}}
Provider used for narrative generation.
Model used for narrative generation
Creativity level: 0=deterministic, 2=creative (default: 0.8)
Minimum words (overrides motion-based defaults)
Maximum words (overrides motion-based defaults)
Max Wikipedia sources to include (default: 10)
Previous narratives sent as context (default: 3)
Beats with complete text for continuity (default: 3)
Older location names to avoid repetition (default: 50)
Narrative continuity instructions injected into the story user prompt.
Transition/opening instructions injected into the story user prompt.
3

Audio Generation

Settings that affect the TTS output

Audio Prompt Layer

Audio generation sends the narrative text directly to TTS.
There is no separate prompt/rewriter stage in phase 3.

Voice Options

VOICE OPTIONS:
- Alloy: Balanced and versatile, good for general use
- Echo: Warm and conversational, slightly lower pitch
- Fable: Expressive and dynamic, British accent
- Onyx: Deep and authoritative (DEFAULT - recommended for narratives)
- Nova: Friendly and upbeat, slightly higher pitch
- Shimmer: Soft and soothing, gentle tone

TTS MODELS:
- tts-1 (Standard): Faster, good quality, lower cost
- tts-1-hd (High Definition): Higher quality audio, 2x cost
- eleven_multilingual_v2: ElevenLabs multilingual TTS
- eleven_turbo_v2_5: ElevenLabs faster TTS

SPEED: 0.25 (very slow) → 1.0 (normal) → 4.0 (very fast)
Choose audio synthesis provider per story.
OpenAI TTS voice (default: Onyx)
Voice ID from ElevenLabs dashboard.
Model list is provider-specific.
OpenAI only: playback speed 0.25-4.0 (default: 1.0)