Skip to content

ADR-006: Context-Aware Model Selection

  • Status: Accepted
  • Date: 2026-01-11
  • Authors: Podcast Scraper Team
  • Related RFCs: RFC-010
  • Related PRDs: PRD-008

Context & Problem Statement

OpenAI Whisper provides generic multilingual models (base, small, etc.) and optimized English-only models (base.en, small.en). The English models are significantly more accurate and faster for English-only audio.

Decision

The system automatically promotes requested models to their English variants if:

  1. The detected or configured language is English (en).
  2. The user requested a model size that has an .en variant (tiny, base, small, medium).
  3. The user hasn't already explicitly specified a variant.

Rationale

  • Quality by Default: Most users want the best accuracy for their language without needing to know technical model suffixes.
  • Resource Efficiency: English models are typically faster and smaller than their multilingual counterparts.

Alternatives Considered

  1. Manual Selection Only: Rejected as it places too much cognitive load on the user.
  2. Multilingual Default: Safe, but leaves significant performance and accuracy on the table for the primary use case (English podcasts).

Consequences

  • Positive: Better transcription quality "out of the box."
  • Negative: Can be surprising to users who explicitly wanted the multilingual model for some reason (e.g., occasional non-English segments).

References