Autonomy Realms

Signal Types


Signals are the core unit of Autonomy Realms. Every signal has a type that determines what file it accepts, how it is processed, and which analysis fields are available.

Transmission

The primary signal type. Transmissions are video or audio recordings — the core use case of the platform.

When you upload a transmission:

  • Video is stored and a thumbnail is auto-generated
  • Audio is extracted from video (or used directly for audio-only files)
  • Transcription runs automatically via OpenAI Whisper, producing a full text transcript
  • The transcript becomes the basis for AI analysis and embedding generation

Transmissions have the richest set of analysis fields, including narrative arc, emotional tone, key quotes, and energy level. This is the type designed for spoken-word recordings, vlogs, interviews, voice memos, and oral history.

Document

Written text content — journal entries, essays, notes, articles, or any prose you want to preserve and analyze.

Documents do not require transcription since the text is the source material. Analysis runs directly on the document body. Use this type for anything you have written rather than spoken.

Photo

Image files with optional EXIF metadata. When you upload a photo:

  • EXIF data is extracted automatically (date taken, GPS coordinates, camera info)
  • Location is populated from GPS coordinates if available
  • The photo appears on the Atlas map if it has location data

Photo analysis fields focus on visual description, setting, mood, and symbolic content rather than transcript-based fields.

Hyperlink

External URLs that you want to reference and preserve within your realm. Hyperlinks point to content hosted elsewhere — articles, videos, resources — and let you annotate and analyze them alongside your own signals.

Analysis Fields by Type

Each signal type has a different set of fields available for analysis. Transcript-dependent fields (like key quotes and narrative arc) are only available for transmissions and other types with text content. Visual fields are emphasized for photos. All types share core fields like title, summary, and tags.

When you select fields for analysis, only the fields relevant to your signal's type are shown.