Gemma 4 cloud audio was infeasible (31b-cloud has no audio; E4B broken
upstream, no deploy host), so improve faster-whisper instead.
- Pin temperature=0.0 to disable the fallback ladder that re-decoded unclear
audio up to 6x (source of the 16-24s latency outliers); reject hallucinated
segments via avg_logprob/compression_ratio in the new pure _filter_segments.
- Adopt mikr/whisper-small-ro-cv11 (CT2 int8) via configurable voice.stt_model:
spike showed WER 24%->10%, numbers fixed at source, +0.33s p50 (in budget).
- Add tools/voice_stt_mine.py (log mining) + tools/voice_stt_spike.py (model
eval with diacritic scoring) + tests for the gate and miner.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Squashed branch: voice/dave-recv → master. Closes Pas 12 (DAVE E2E) and lands
voice-mode UX polish + verbal voice control on top of the Pas 1-10 scaffolding
already on master.
## DAVE E2E receive-side decrypt (e4f3177)
Vendored fork: discord-ext-voice-recv 0.5.3a+echo.dave1. Patches the receive
pipeline to handle Discord's mandatory DAVE encryption on voice gateway v=8.
- `_maybe_dave_decrypt`: uses davey.can_passthrough(user_id) as primary gate,
falls through to dave.decrypt for DAVE-epoch peers, drops on decrypt failure
without killing the reader thread.
- VAD fix: silero-vad v5+ requires exactly 512 samples; our 100ms window
(1600 samples) was silently raising ValueError → STT never fired. Now slice
into 512-sample chunks.
- Whisper: bumped beam_size 1→5 and added RO initial_prompt.
- Tests: 11 DAVE unit tests + 2 callback integration tests + contract test
with fork-version guard.
## Voice UX polish (d1bc77e)
- Killed the 3s "mă gândesc" filler (always collided with Claude p50 4-7s).
- Barge-in via `ttsq.clear()` at top of `on_segment_done`.
- DTX silence-flush poller (200ms tick) — Discord stops sending RTP packets
when silent, so the inline silence-check in sink.write() never fired for
trailing audio; background thread handles it.
- `EchoStreamingAudioSource.read()` non-blocking — old `get_frame(timeout=0.1)`
wrecked Discord's 20ms cadence and the client interpreted bursts as
stuttering (Marius heard "4 de minute" instead of full sentence).
- RO time expansion: 23:09 → "douăzeci și trei și nouă minute".
- Supertonic Unicode sanitize centralized in tools/tts.py.
- Whisper local_files_only=True — no HF metadata GET on each startup.
- Diagnostic logging through sink → VAD → Claude stream → TTS chain.
## Voice mode iteration (e589e48)
- `personality/VOICE_MODE.md` — voice-tailored system prompt (short, no
markdown, no abbreviations, time without seconds, distances in
"mii"/"milioane"); plumbed via build_system_prompt(voice_mode=True).
- Isolated voice session key `voice:<channel_id>` — voice doesn't share
context with text adapter on the same channel; auto-applied without
/clear ceremony. /clear drops both keys.
- Metric units + Romanian thousands (normalize.py): "384.000 km" →
"trei sute optzeci și patru de mii de kilometri" with feminine-correct
pluralization and "de" particle for ≥20.
- `/voice setvoice <M1-F5>` slash command with native autocomplete; swaps
live + persists voice.default_voice to config.json.
- Verbal voice change (src/voice/voice_commands.py + 29 tests) — "schimbă
vocea pe M5", "voce em cinci", with permissive substring fallback for
Whisper-mangled forms like "Mâcinci"=M5 and "unul cinci"=M5. Whisper
initial_prompt now lists voice vocabulary to bias STT toward clean
outputs.
- Fast barge-in: VAD ≥2 consecutive windows (~200ms) on Marius's user
while Echo has pending TTS frames → cut him off mid-sentence so user
doesn't wait the full silence + STT cycle. Acoustic echo bleed-through
still requires headphones (no AEC).
## Test suite
130 voice + router tests pass (test_voice_recv_dave, test_voice_session_cleanup,
test_voice_adapter_contract, test_voice_normalize, test_voice_commands,
test_router).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Imported claude jobs default channel to echo-work; grup-sprijin-5feb
and grup-sprijin-pregatire route to echo-sprijin. Existing echo-core
channel is preserved.
Loop through consecutive newsletter numbers until one is missing, so
backlog gets delivered in a single run. Use httpx for 404 check and
point to absolute claude binary path for cron. Enable job in config.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- heartbeat saves unread whitelisted emails via email_process --save --json
- fix: add --add-dir so Claude CLI subprocess can access memory/ symlink
- email_check/process: use BODY.PEEK[] to avoid marking emails as read
- email_process: simplify credential loading via credential_store only
- config: heartbeat interval 30→120min, quiet hours end 08→07
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Switch Bash permission patterns from space to colon separator
- Add memory.bak/ to .gitignore
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Fast commands for git, email, calendar, notes, search, reminders, and
diagnostics — all execute instantly without Claude CLI. Incremental
embeddings indexing in heartbeat (1h cooldown) + inline indexing after
/note, /jurnal, /email save. Fix Ollama URL (localhost → 10.0.20.161),
fix email_process.py KB path (kb/ → memory/kb/).
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Calendar no longer bypasses quiet hours. First run after quiet hours
sends full daily summary, subsequent runs only remind for next event
within 45 min with deduplication. Calendar cooldown set to 30 min.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Heartbeat system overhaul:
- Fix email/calendar checks to parse JSON output correctly
- Add per-check cooldowns and quiet hours config
- Send findings to Discord channel instead of just logging
- Auto-reindex KB when stale files detected
- Claude CLI called only if HEARTBEAT.md has extra instructions
- All settings configurable via config.json heartbeat section
Move hardcoded values to config.json:
- allowed_tools list (claude_session.py)
- Ollama URL/model (memory_search.py now reads ollama.url from config)
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Bridge: allow fromMe messages in groups, include participant field in
message queue, bind to 0.0.0.0 for network access, QR served as HTML.
Adapter: process registered group messages (route to Claude), extract
participant for user identification, fix unbound 'phone' variable.
Tested end-to-end: WhatsApp group chat with Claude working. 442 tests pass.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Node.js bridge (bridge/whatsapp/): Baileys client with Express HTTP API
on localhost:8098 — QR code linking, message queue, reconnection logic.
Python adapter (src/adapters/whatsapp.py): polls bridge every 2s, routes
through router.py, separate whatsapp.owner/admins auth, security logging.
Integrated in main.py alongside Discord + Telegram via asyncio.gather.
CLI: echo whatsapp status/qr. 442 tests pass (32 new, zero failures).
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
