hermes/svrnty-hermes-webui-plugin

Fork 0

Svrnty plugin for nesquena/hermes-webui — single repo for every backend + brand mod (per hermes/docs/SVRNTY-PLUGIN-PROTOCOL.md)

Go to file

Svrnty 0ef66ab599 All checks were successful plugin-tests / test (push) Successful in 6s Details feat(plugin): voice-message mic UI in app.js — closes Phase 2.A UX gap (L8) Migrates the boot.js mic-button behavior change from reverted fork commit 014b9eef into plugin/static/app.js. Replaces vanilla dictation-to-textbox mic flow with voice-message mode: tap to record, tap again to stop → audio File attached + sent automatically. Server-side audio_attachment_processor in routes/transcribe.py transcribes for text-only models. Implementation strategy: - Own #btnMic onclick (boot.js sets it during init; plugin overrides via MutationObserver after DOM mutations + idempotent dataset.svrntyVm flag) - MediaRecorder with same mime-type fallbacks as the original (webm/opus → webm → ogg/opus → mp4) - DataTransfer to set #fileInput.files programmatically (cross-browser path) - Dispatch 'change' so boot.js's fileInput.onchange → addFiles() runs - Click #btnSend after 50ms tick so attachment registers before submit Filename convention: voice-message-${timestamp}.{webm\|ogg\|mp4} — matches the audio processor's name-prefix detection in _transcribe_audio_attachments. Tests: 6 new JS-static checks in tests/unit/test_app_js.py (idempotent guard, voice-message prefix, DataTransfer pattern, vault URL pin, mic override). 26 tests total, all PASS. Connection map: now 9 public API · 0 forced internal · 1 frontend. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>		2026-05-23 10:28:47 -04:00
.github/workflows	feat(plugin): initial scaffold — plugin loader entry + AST + CI workflows (P1.B/C/D)	2026-05-23 09:59:45 -04:00
routes	feat(plugin): STT migration via audio_attachment_processor hook (L1-L6)	2026-05-23 10:14:29 -04:00
scripts	feat(plugin): STT migration via audio_attachment_processor hook (L1-L6)	2026-05-23 10:14:29 -04:00
static	feat(plugin): voice-message mic UI in app.js — closes Phase 2.A UX gap (L8)	2026-05-23 10:28:47 -04:00
svrnty_hermes_webui_plugin	feat(plugin): initial scaffold — plugin loader entry + AST + CI workflows (P1.B/C/D)	2026-05-23 09:59:45 -04:00
tests	feat(plugin): voice-message mic UI in app.js — closes Phase 2.A UX gap (L8)	2026-05-23 10:28:47 -04:00
.gitignore	feat(plugin): initial scaffold — plugin loader entry + AST + CI workflows (P1.B/C/D)	2026-05-23 09:59:45 -04:00
CLAUDE.md	feat(plugin): initial scaffold — plugin loader entry + AST + CI workflows (P1.B/C/D)	2026-05-23 09:59:45 -04:00
CONNECTION-MAP.md	feat(plugin): voice-message mic UI in app.js — closes Phase 2.A UX gap (L8)	2026-05-23 10:28:47 -04:00
Makefile	feat(plugin): Phase 2 partial — vault_status migrated + brand skin moved + eval suite (P2.B/C, P3.A/B)	2026-05-23 10:02:47 -04:00
manifest.yaml	feat(plugin): initial scaffold — plugin loader entry + AST + CI workflows (P1.B/C/D)	2026-05-23 09:59:45 -04:00
plugin.py	feat(plugin): STT migration via audio_attachment_processor hook (L1-L6)	2026-05-23 10:14:29 -04:00
pyproject.toml	feat(plugin): initial scaffold — plugin loader entry + AST + CI workflows (P1.B/C/D)	2026-05-23 09:59:45 -04:00
README.md	feat(plugin): initial scaffold — plugin loader entry + AST + CI workflows (P1.B/C/D)	2026-05-23 09:59:45 -04:00

README.md

svrnty-hermes-webui-plugin

THE single repo holding every Svrnty modification to nesquena/hermes-webui. Loaded at runtime via the plugin loader hook patched into the fork.

Protocol contract: hermes/docs/SVRNTY-PLUGIN-PROTOCOL.md — read this before contributing.

Install

# 1. Install the plugin in the same venv as hermes-webui
pip install -e ~/workspaces/hermes/svrnty-hermes-webui-plugin

# 2. Tell hermes-webui to load it
export HERMES_WEBUI_PYTHON_PLUGIN=svrnty_hermes_webui_plugin
# or set in docker-compose.override.yml under environment:

# 3. Restart hermes-webui — endpoints under /api/* + assets under /plugins/svrnty/* land

Without HERMES_WEBUI_PYTHON_PLUGIN, hermes-webui runs vanilla (no Svrnty mods).

What's in here

Dir	What
`plugin.py`	Entry point — `register(api)` wires routes + static
`routes/`	One file per Svrnty `/api/*` endpoint (transcribe, vault_status, …)
`static/`	Brand skin: `app.js`, `app.css`, Montserrat fonts
`CONNECTION-MAP.md`	AST-generated map of every upstream symbol this plugin touches
`scripts/`	Tooling — AST walker, upstream sync, boot smoke
`tests/`	Unit · integration · evals (each upstream-sync runs evals)
`.github/workflows/`	Plugin-tests · connection-map-check · upstream-drift CI

Public extension API

The plugin loader (one fork commit in hermes-webui) exposes exactly 6 methods:

api.register_route(path, method, handler)   # add /api/<path>
api.register_static(prefix, directory)      # serve files under /plugins/<prefix>/...
api.inject_script(url)                      # add <script> to index.html
api.inject_stylesheet(url)                  # add <link> to index.html
api.config_get(key, default)                # safe upstream config read
api.logger(name)                            # namespaced logger

Touching anything else in hermes-webui = a Rule 2 violation per the protocol. Document the escape hatch in CONNECTION-MAP.md under "forced internal dependencies".

Hygiene

make sync-upstream — one-command rebase against latest upstream + report
python scripts/ast-connection-map.py — regenerate the map
python scripts/boot-smoke.py — start + curl every plugin endpoint
pytest tests/ — full suite (unit + integration + evals)

Status

Component	State
Loader hook in `hermes-webui`	TBD (Phase 1)
Plugin scaffold	Phase 1 in progress
Migrated features (transcribe, vault_status, brand skin)	TBD (Phase 2)
Automation (drift CI, sync command, eval suite)	TBD (Phase 3)
Upstream PR	TBD (Phase 4)