DeepGram Speech platform

**Judgment:** A strong, focused CLI wrapper for Deepgram’s speech-to-text API. Well-suited for developers and power users who want fast, scriptable transcription from various audio sources without building their own UI. **What it does well:** - Handles local files, URLs, stdin, and microphone input with a single, consistent `deepgram listen` interface. - Exposes key Deepgram options (model, language, punctuation, diarization, output format) so you can tune behavior and integrate results into downstream tools. - CLI-first design makes it easy to embed in shell scripts, data pipelines, and automated workflows, especially when dealing with JSON or subtitle formats. **Risks / Limitations:** - **Privacy & compliance:** All audio is sent to Deepgram’s cloud; this may be unsuitable for highly sensitive, regulated, or air-gapped environments. Users must ensure proper consent and data handling. - **External dependency:** Reliability and latency depend on network connectivity and Deepgram’s service availability and rate limits. - **Access & cost:** Requires a Deepgram API key and is subject to Deepgram pricing and quota constraints. **Recommended scenarios:** - You want a **scriptable speech-to-text tool** for batch transcription, log enrichment, or content indexing. - You are building **automation or pipelines** that take audio (files, URLs, or streams), transcribe it, then search, summarize, or generate subtitles. - You need **quick, no-UI live transcription** from a microphone for demos, experiments, or feeding into other agents or tools. Avoid this skill if you absolutely cannot send audio to third-party cloud services or need fine-grained on-prem model control.