From 17fbb7c1dd7975108d5cd5fe2ecb5fee53279ec3 Mon Sep 17 00:00:00 2001 From: Bennie Williams Date: Sat, 31 Oct 2026 10:06:42 -0400 Subject: [PATCH] feat(catalog): add breeze-tts, a Breeze Blue TTS provider MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Pins MDsniper/hermes-breeze-tts at 19fda86: a standalone plugin that registers the 'breeze' TTSProvider against the Breeze Blue Developer API. Catalog metadata only — no core code touched. --- plugin-catalog/breeze-tts.yaml | 24 ++++++++++++++++++++++++ 1 file changed, 24 insertions(+) create mode 100644 plugin-catalog/breeze-tts.yaml diff --git a/plugin-catalog/breeze-tts.yaml b/plugin-catalog/breeze-tts.yaml new file mode 100644 index 0000000000..14936d23f4 --- /dev/null +++ b/plugin-catalog/breeze-tts.yaml @@ -0,0 +1,24 @@ +name: breeze-tts +title: Breeze Blue TTS +repo: https://github.com/MDsniper/hermes-breeze-tts +subdir: breeze-tts +sha: a7b9e6ffa0143d7f07562d8708fd16a2f10eb2fb +description: "Breeze Blue text-to-speech as a Hermes voice provider (tts.provider: breeze-blue), driven by the Breeze Blue Developer API. Voices come from the Breeze console — stock, cloned or designed — and the provider speaks whole replies through text_to_speech and read-aloud, with mp3 (default), wav, flac, opus, aac or pcm output, so Opus voice notes work on messaging platforms. No new Python dependencies: the plugin posts directly to the ElevenLabs-shaped endpoint. Disclosure — sends the text being spoken, the chosen voice id and model, and your BREEZE_API_KEY (header xi-api-key) to api.breeze.blue, or to BREEZE_BASE_URL when you set it; synthesis is billed to your Breeze Blue account; generated audio is written to Hermes's audio cache; the text is not logged; no tools, hooks or middleware; no telemetry. Not implemented: realtime/WebSocket streaming (voice-mode replies are spoken sentence by sentence instead) and voice listing." +maintainer: MDsniper +tier: community +category: voice +requires_hermes: ">=0.21.6" +docs_url: https://github.com/MDsniper/hermes-breeze-tts#readme +version: "0.1.0" +readme: true +platforms: [] +known_issues: + - "The provider is named breeze-blue, not breeze: 'breeze' is the self-hosted local streaming server keyed under tts.streaming.provider (PR #105299), so the hosted API provider takes its vendor name." + - "tts.voice is mandatory: Breeze Blue has no built-in voice and every voice id is account-scoped, so the provider raises instead of guessing one." + - "No voice listing or realtime streaming; per-sentence synthesis covers voice mode at higher latency." +capabilities: + provides_tools: [] + provides_hooks: [] + provides_middleware: [] + requires_env: + - BREEZE_API_KEY