Deepgram

Harvey — Runs a solar company's CRM, mail, calendar, documents, bookkeeping and equipment monitoring, with an AI assistant that can act on them by chat or voice.

Afterword — Generates structured session notes, follow-up emails, and client prep briefs from coaching call transcripts or recordings.

RayMate — Provides a Mac launcher panel for AI chat, on-device or cloud dictation, and inline text rewriting with diff review.

Clerune — Turns meeting recordings or notes into minutes, action items, and follow-up tasks.

Mock Interview Copilot — Listens to a mock interview call's tab audio, detects the interviewer's question, and suggests short answers drawn from the user's uploaded notes.

Kommz — Translates voice speech in real time during online gaming so players can talk across languages.

Kommz Gamer — Translates players' voice chat in real time between languages and routes it through a virtual audio device.

Hail — Provides a self-hosted server that handles email, SMS, and voice phone bot communication for software projects, exposed via API, MCP, CLI, and Python SDK.

Voice of Dragons — Runs a voice-only tabletop role-playing campaign where one to four players speak their actions and an AI game master narrates the story.

Paxio — Runs a personal AI agent over WhatsApp that accepts text or voice queries and returns streamed step updates and summaries.

monkeyspeak — Scores timed spoken prompts for speaking speed and clarity, using speech-to-text transcripts compared against the prompt.

PodTLDR.fm — Transcribes and summarizes new episodes of followed podcasts and emails a TLDR with takeaways, quotes, and a verdict on whether to listen.

Notai — Records audio and produces a transcript with speaker labels, live transcription, an AI summary with action items and sentiment, and chat over the notes.

Pipevoice — Transcribes held-hotkey speech and types the resulting text as keystrokes into the focused window, with optional AI cleanup of filler words.

livekit-starter — Provides a full-stack starter template for building real-time voice agents with a LiveKit voice worker, a FastAPI token server, and a React frontend.

sarvam-bridge — Proxies ElevenLabs, OpenAI, and Deepgram style speech API calls to Sarvam AI, handling base64 decoding, language codes, text chunking, and speaker selection.