psd-tts

name	psd-tts
summary	Turn text into a shareable MP3 with Amazon Polly (generative voices) and return a public HTTPS link — narration, briefings, audio summaries.
description	Convert text to natural-sounding speech (MP3) with Amazon Polly and return a shareable HTTPS URL. Use when asked to read something aloud, narrate a document, or make an audio version or podcast of text.
allowed-tools	Bash(/opt/agentcore-venv/bin/python3:*)

psd-tts

Convert text to a natural-sounding MP3 with Amazon Polly, upload it to the workspace S3 bucket under the public public-images/ prefix, and return an unsigned HTTPS URL the user can play in any browser (Google Chat renders it as a link). Same delivery model as psd-image-gen and psd-html-artifact.

Identity. Requires --user <caller-email>. Pass the email verbatim from the [caller: Name <email>] header of the user turn (it scopes the S3 upload path).

Polly is a standard AWS service (not Bedrock) authenticated by the execution role — no API key, no per-character prompt to the model. Long text is split at sentence boundaries and the MP3 chunks are concatenated automatically.

Usage

Text comes from --text, --file, or stdin (safe for long input):

/opt/agentcore-venv/bin/python3 /opt/psd-skills/psd-tts/scripts/synthesize.py --user <email> --text "Good morning. Here is today's briefing."

printf '%s' "$LONG_TEXT" | /opt/agentcore-venv/bin/python3 /opt/psd-skills/psd-tts/scripts/synthesize.py --user <email> --voice Matthew

Options:

Flag	Default	Description
--voice	Ruth	Polly voice id (see references/voices.md)
--engine	generative	generative | neural | long-form | standard
--file	—	Read text from a container file instead of --text

Returns JSON: { "status": "ok", "url": "...", "s3Key": "public-images/.../<uuid>.mp3", "voice": "Ruth", "engine": "generative", "characters": N, "chunks": N, "sharing": "public-by-link" }.

Required Reply Format

After the skill returns status: ok, your next chat message MUST contain the bare url on a line by itself (Google Chat renders it as a playable/downloadable link). Do not wrap it in Markdown link syntax or bold — bare URL only. The URL is public-by-link and does not expire; do not presign or re-upload it.

Voice & engine guidance

• generative (default) is Polly's most natural, expressive engine. Default voice Ruth (warm en-US female); Matthew is the male counterpart.
• long-form is tuned for narrating articles/training material (en-US: Danielle, Gregory, Patrick, Ruth) — reach for it on long documents.
• neural has the widest voice/language coverage if you need a voice the generative engine doesn't offer.
• Generative voices are region-limited but supported in this deployment's region (us-east-1). See references/voices.md for the full list before overriding the default.

Errors

• bad_args — missing/invalid --user, or no text supplied.
• too_large — text exceeds 200,000 characters.
• upstream_error — Polly rejected the request (e.g. a voice not offered by the chosen engine) or the owner-bound artifact broker failed. Check the voice/engine pairing in references/voices.md.
• misconfigured — the trusted web broker is not configured.