PennyScribe is an API‑based audio transcription service designed for developers and AI agents. It converts spoken content from interviews, podcasts, sales calls, and lectures into clean, plain text without timestamps or extraneous formatting. The service accepts audio files up to 5 hours in length and supports transcription in over 32 languages, including varied dialects and challenging acoustic environments such as rapid speech, singing, and background noise.
Key capabilities:
RESTful API with key‑based authentication and webhook support for asynchronous result delivery.
No minimum file size or duration limits for single submissions (up to 5 hours per file).
Transcriptions are returned as unadorned text, suitable for direct consumption by LLMs, search indexes, or custom processing pipelines.
Uploaded audio files are not retained after transcription, ensuring data privacy.
Pricing model:
Pay‑as‑you‑go, with no monthly subscription or long‑term commitment. Currently, 2,400 minutes of transcription cost $5 (approximately $0.0021 per minute). Billing applies only to successfully completed tasks; failed transcriptions are not charged.
Intended use cases:
Research & analysis – transcribe client interviews or focus groups for qualitative coding and summarisation.
Content repurposing – generate written show notes, newsletters, or social media drafts from podcast episodes.
Sales intelligence – extract patterns and insights from discovery calls and customer conversations.
Automated ingestion – integrate speech‑to‑text into larger data workflows, with results delivered via webhook without keeping browsers waiting.
The service is built to function as a foundational speech‑to‑text layer within AI agent architectures and batch processing systems. All features and pricing are current as of August 2026.
