plugin

saymd

Voice to Markdown prompts

saymd turns speech or audio files into a structured Markdown prompt (Objective, Context, Instructions, Constraints) that Codex can @. Runs your local Free MIT CLI (npx saymd) with Bring Your Own Key for Gemini, OpenAI, Deepgram, or ElevenLabs. Audio never goes to saymd.app.

Typical loop: record with the CLI → open .ai/prompt.md → implement from the spec.

Install

  1. 1Add it in ChatGPT

    Open Plugins in the ChatGPT sidebar (web or desktop), search for “saymd” and select +. You'll be asked to sign in to any connected service on first use. Then type @ in a chat to call it.

Skills 1

  • saymd

    Turn spoken ideas or audio files into a structured Markdown prompt file for coding agents. Use when the user wants to dictate a feature, bug, or plan; record a voice note into a reusable agent prompt; or run saymd / saymd --file. Prefer the local CLI over in-chat transcription. After recording, always open or @ .ai/prompt.md and offer to implement.

Things to ask

  • “Use saymd to record a feature idea into .ai/prompt.md, then implement from that file.”
  • “Turn a voice bug report into .ai/prompt.md with saymd, then fix from the structured spec.”
  • “Run saymd with the plan template and start coding from .ai/prompt.md.”

Related