When the Discord typing API call fails (rate limit, network error, 403),
_typing_loop returns early but the stale task remains in _typing_tasks.
Subsequent send_typing calls see the stale entry and skip, leaving no
typing indicator for the rest of the agent invocation.
Add finally block to _typing_loop to always remove the task from
_typing_tasks on exit, whether from cancellation, error, or normal
completion. This allows send_typing to create a fresh task.
3 new tests in test_discord_send.py:
- Task removed after API error
- Typing restartable after failure
- stop_typing cleans up
1. Quick command exec ran in the gateway process's full environment
without env sanitization or output redaction. A quick command like
"env" or "printenv" would leak all API keys, OAuth tokens, and
bot credentials to the messaging user.
Fix: apply _sanitize_subprocess_env() before exec and
redact_sensitive_text() on output before returning.
2. GatewayRunner._pending_messages was written on every interrupt
(lines 1331-1334) but never read or consumed anywhere. The actual
interrupt delivery uses adapter._pending_messages (a separate dict).
Removed the write-only accumulation to prevent unbounded growth.
`_normalize_for_deepseek` was mapping every non-reasoner input into
`deepseek-chat` on the assumption that DeepSeek's API accepts only two
model IDs. That assumption no longer holds — `deepseek-v4-pro` and
`deepseek-v4-flash` are first-class IDs accepted by the direct API,
and on aggregators `deepseek-chat` routes explicitly to V3 (DeepInfra
backend returns `deepseek-chat-v3`). So a user picking V4 Pro through
the model picker was being silently downgraded to V3.
Verified 2026-04-24 against Nous portal's OpenAI-compat surface:
- `deepseek/deepseek-v4-flash` → provider: DeepSeek,
model: deepseek-v4-flash-20260423
- `deepseek/deepseek-chat` → provider: DeepInfra,
model: deepseek/deepseek-chat-v3
Fix:
- Add `deepseek-v4-pro` and `deepseek-v4-flash` to
`_DEEPSEEK_CANONICAL_MODELS` so exact matches pass through.
- Add `_DEEPSEEK_V_SERIES_RE` (`^deepseek-v\d+(...)?$`) so future
V-series IDs (`deepseek-v5-*`, dated variants) keep passing through
without another code change.
- Update docstring + module header to reflect the new rule.
Tests:
- New `TestDeepseekVSeriesPassThrough` — 8 parametrized cases covering
bare, vendor-prefixed, case-variant, dated, and future V-series IDs
plus end-to-end `normalize_model_for_provider(..., "deepseek")`.
- New `TestDeepseekCanonicalAndReasonerMapping` — regression coverage
for canonical pass-through, reasoner-keyword folding, and
fall-back-to-chat behaviour.
- 77/77 pass.
Reported on Discord (Ufonik, Don Piedro): `/model > Deepseek >
deepseek-v4-pro` surfaced
`Normalized 'deepseek-v4-pro' to 'deepseek-chat'`. Picker listing
showed the v4 names, so validation also rejected the post-normalize
`deepseek-chat` as "not in provider listing" — the contradiction
users saw. Normalizer now respects the picker's choice.
- hermes_cli/auth.py: add _default_verify() with macOS Homebrew certifi
fallback (mirrors weixin 3a0ec1d93). Extend env var chain to include
REQUESTS_CA_BUNDLE so one env var works across httpx + requests paths.
- agent/model_metadata.py: add _resolve_requests_verify() reading
HERMES_CA_BUNDLE / REQUESTS_CA_BUNDLE / SSL_CERT_FILE in priority
order. Apply explicit verify= to all 6 requests.get callsites.
- Tests: 18 new unit tests + autouse platform pin on existing
TestResolveVerifyFallback to keep its "returns True" assertions
platform-independent.
Empirically verified against self-signed HTTPS server: requests honors
REQUESTS_CA_BUNDLE only; httpx honors SSL_CERT_FILE only. Hermes now
honors all three everywhere.
Triggered by Discord reports — Nous OAuth SSL failure on macOS
Homebrew Python; custom provider self-signed cert ignored despite
REQUESTS_CA_BUNDLE set in env.
The voice.toggle handler was persisting display.voice_enabled /
display.voice_tts to config.yaml, so a TUI session that ever turned
voice on would re-open with it already on (and the mic badge lit) on
every subsequent launch. cli.py treats voice strictly as runtime
state: _voice_mode = False at __init__, only /voice on flips it, and
nothing writes it back to disk.
Drop the _write_config_key calls in voice.toggle on/off/tts and the
config.yaml fallback in _voice_mode_enabled / _voice_tts_enabled.
State is now env-var-only (HERMES_VOICE / HERMES_VOICE_TTS), scoped to
the live gateway subprocess — the next launch starts clean.
Crash-log stack trace (tui_gateway_crash.log) from the user's session
pinned the regression: SIGPIPE arrived while main thread was blocked on
for-raw-in-sys.stdin — i.e., a background thread (debug print to stderr,
most likely from HERMES_VOICE_DEBUG=1) wrote to a pipe whose buffer the
TUI hadn't drained yet, and SIG_DFL promptly killed the process.
Two fixes that together restore CLI parity:
- entry.py: SIGPIPE → SIG_IGN instead of the _log_signal handler that
then exited. With SIG_IGN, Python raises BrokenPipeError on the
offending write, which write_json already handles with a clean exit
via _log_exit. SIGTERM / SIGHUP still route through _log_signal so
real termination signals remain diagnosable.
- hermes_cli/voice.py:_debug: wrap the stderr print in a BrokenPipeError
/ OSError try/except. This runs from daemon threads (silence callback,
TTS playback, beep), so a broken stderr must not escape and ride up
into the main event loop.
Verified by spawning the gateway subprocess locally:
voice.toggle status → 200 OK, process stays alive, clean exit on
stdin close logs "reason=stdin EOF" instead of a silent reap.
SIG_DFL for SIGPIPE means the kernel reaps the gateway subprocess the
instant a background thread (TTS playback, silence callback, voice
status emitter) writes to a stdout the TUI stopped reading — before
the Python interpreter can run excepthook, threading.excepthook,
atexit, or the entry.py post-loop _log_exit.
Replace the three SIG_DFL / SIG_IGN bindings with a _log_signal
handler that:
- records which signal (SIGPIPE / SIGTERM / SIGHUP) fired and when;
- dumps the main-thread stack at signal delivery AND every live
thread's stack via sys._current_frames — the background-thread
write that provoked SIGPIPE is almost always visible here;
- writes everything to ~/.hermes/logs/tui_gateway_crash.log and prints
a [gateway-signal] breadcrumb to stderr so the TUI Activity surfaces
it as well.
SIGINT stays ignored (TUI handles Ctrl+C for the user).
Gateway exits weren't reaching the panic hook because entry.py calls
sys.exit(0) on broken stdout — clean termination, no exception. That
left "gateway exited" in the TUI with zero forensic trail when pipe
breaks happened mid-turn.
Entry.py now tags each exit path — startup-write failure, parse-error-
response write failure, per-method response write failure, stdin EOF —
with a one-line entry in ~/.hermes/logs/tui_gateway_crash.log and a
gateway.stderr breadcrumb. Includes the JSON-RPC method name on the
dispatch path, which is the only way to tell "died right after handling
voice.toggle on" from "died emitting the second message.complete".
When the gateway subprocess raises an unhandled exception during a
voice-mode turn, nothing survives: stdout is the JSON-RPC pipe, stderr
flushes but the process is already exiting, and no log file catches
Python's default traceback print. The user is left with an
undiagnosable "gateway exited" banner.
Install:
- sys.excepthook → write full traceback to tui_gateway_crash.log +
echo the first line to stderr (which the TUI pumps into
Activity as a gateway.stderr event). Chains to the default hook so
the process still terminates.
- threading.excepthook → same, tagged with the thread name so it's
clear when the crash came from a daemon thread (beep playback, TTS,
silence callback, etc.).
- Turn-dispatcher except block now also appends a traceback to the
crash log before emitting the user-visible error event — str(e)
alone was too terse to identify where in the voice pipeline the
failure happened.
Zero behavioural change on the happy path; purely forensics.
TTS feedback loop (hermes_cli/voice.py)
The VAD loop kept the microphone live while speak_text played the
agent's reply over the speakers, so the reply itself was picked up,
transcribed, and submitted — the agent then replied to its own echo
("Ha, looks like we're in a loop").
Ported cli.py:_voice_tts_done synchronisation:
- _tts_playing: threading.Event (initially set = "not playing").
- speak_text cancels the active recorder before opening the speakers,
clears _tts_playing, and on exit waits 300 ms before re-starting the
recorder — long enough for the OS audio device to settle so afplay
and sounddevice don't race for it.
- _continuous_on_silence now waits on _tts_playing (up to 60 s) before
re-arming the mic with another 300 ms gap, mirroring
cli.py:10619-10621. If the user flips voice off during the wait the
loop exits cleanly instead of fighting for the device.
Without both halves the loop races: if the silence callback fires
before TTS starts it re-arms immediately; if TTS is already playing
the pause-and-resume path catches it.
Red REC badge (ui-tui appChrome + useMainApp)
Classic CLI (cli.py:_get_voice_status_fragments) renders "● REC" in
red and "◉ STT" in amber. TUI was showing a dim "REC" with no dot,
making it hard to spot at a glance. voiceLabel now emits the same
glyphs and appChrome colours them via t.color.error / t.color.warn,
falling back to dim for the idle label.
Three issues surfaced during end-to-end testing of the CLI-parity voice
loop and are fixed together because they all blocked "speak → agent
responds → TTS reads it back" from working at all:
1. Wrong result key (hermes_cli/voice.py)
transcribe_recording() returns {"success": bool, "transcript": str},
matching cli.py:_voice_stop_and_transcribe. The wrapper was reading
result.get("text"), which is None, so every successful Groq / local
STT response was thrown away and the 3-strikes halt fired after
three silent-looking cycles. Fixed by reading "transcript" and also
honouring "success" like the CLI does. Updated the loop simulation
tests to return the correct shape.
2. TTS speak-back was missing (tui_gateway/server.py + hermes_cli/voice.py)
The TUI had a voice.toggle "tts" subcommand but nothing downstream
actually read the flag — agent replies never spoke. Mirrored
cli.py:8747-8754's dispatch: on message.complete with status ==
"complete", if _voice_tts_enabled() is true, spawn a daemon thread
running speak_text(response). Rewrote speak_text as a full port of
cli.py:_voice_speak_response — same markdown-strip regex pipeline
(code blocks, links, bold/italic, inline code, headers, list bullets,
horizontal rules, excessive newlines), same 4000-char cap, same
explicit mp3 output path, same MP3-over-OGG playback choice (afplay
misbehaves on OGG), same cleanup of both extensions. Keeps TUI TTS
audible output byte-for-byte identical to the classic CLI.
3. Auto-submit swallowed on non-empty composer (createGatewayEventHandler.ts)
The voice.transcript handler branched on prev input via a setInput
updater and fired submitRef.current inside the updater when prev was
empty. React strict mode double-invokes state updaters, which would
queue the submit twice; and when the composer had any content the
transcript was merely appended — the agent never saw it. CLI
_pending_input.put(transcript) unconditionally feeds the transcript
as the next turn, so match that: always clear the composer and
setTimeout(() => submitRef.current(text), 0) outside any updater.
Side effect can't run twice this way, and a half-typed draft on the
rare occasion is a fair trade vs. silently dropping the turn.
Also added peak_rms to the rec.stop debug line so "recording too quiet"
is diagnosable at a glance when HERMES_VOICE_DEBUG=1.
The TUI had drifted from the CLI's voice model in two ways:
- /voice on was lighting up the microphone immediately and Ctrl+B was
interpreted as a mode toggle. The CLI separates the two: /voice on
just flips the umbrella bit, recording only starts once the user
presses Ctrl+B, which also sets _voice_continuous so the VAD loop
auto-restarts until the user presses Ctrl+B again or three silent
cycles pass.
- /voice tts was missing entirely, so users couldn't turn agent reply
speech on/off from inside the TUI.
This commit brings the TUI to parity.
Python
- hermes_cli/voice.py: continuous-mode API (start_continuous,
stop_continuous, is_continuous_active) layered on the existing PTT
wrappers. The silence callback transcribes, fires on_transcript,
tracks consecutive no-speech cycles, and auto-restarts — mirroring
cli.py:_voice_stop_and_transcribe + _restart_recording.
- tui_gateway/server.py:
- voice.toggle now supports on / off / tts / status. The umbrella
bit lives in HERMES_VOICE + display.voice_enabled; tts lives in
HERMES_VOICE_TTS + display.voice_tts. /voice off also tears down
any active continuous loop so a toggle-off really releases the
microphone.
- voice.record start/stop now drives start_continuous/stop_continuous.
start is refused with a clear error when the mode is off, matching
cli.py:handle_voice_record's early return on `not _voice_mode`.
- New voice.transcript / voice.status events emit through
_voice_emit (remembers the sid that last enabled the mode so
events land in the right session).
TypeScript
- gatewayTypes.ts: voice.status + voice.transcript event
discriminants; VoiceToggleResponse gains tts; VoiceRecordResponse
gains status for the new "started/stopped" responses.
- interfaces.ts: GatewayEventHandlerContext gains composer.setInput +
submission.submitRef + voice.{setRecording, setProcessing,
setVoiceEnabled}; InputHandlerContext.voice gains enabled +
setVoiceEnabled for the mode-aware Ctrl+B handler.
- createGatewayEventHandler.ts: voice.status drives REC/STT badges;
voice.transcript auto-submits when the composer is empty (CLI
_pending_input.put parity) and appends when a draft is in flight.
no_speech_limit flips voice off + sys line.
- useInputHandlers.ts: Ctrl+B now calls voice.record (start/stop),
not voice.toggle, and nudges the user with a sys line when the
mode is off instead of silently flipping it on.
- useMainApp.ts: wires the new event-handler context fields.
- slash/commands/session.ts: /voice handles on / off / tts / status
with CLI-matching output ("voice: mode on · tts off").
Backward compat preserved for voice.record (was always PTT shape;
gateway still honours start/stop with mode-gating added).
tui_gateway/server.py:3486/3491/3509 imports start_recording,
stop_and_transcribe, and speak_text from hermes_cli.voice, but the
module never existed (not in git history — never shipped, never
deleted). Every voice.record / voice.tts RPC call hit the ImportError
branch and the TUI surfaced it as "voice module not available — install
audio dependencies" even on boxes with sounddevice / faster-whisper /
numpy installed.
Adds a thin wrapper on top of tools.voice_mode (recording +
transcription) and tools.tts_tool (text-to-speech):
- start_recording() — idempotent; stores the active AudioRecorder in a
module-global guarded by a Lock so repeat Ctrl+B presses don't fight
over the mic.
- stop_and_transcribe() — returns None for no-op / no-speech /
Whisper-hallucination cases so the TUI's existing "no speech detected"
path keeps working unchanged.
- speak_text(text) — lazily imports tts_tool (optional provider SDKs
stay unloaded until the first /voice tts call), parses the tool's
JSON result, and plays the audio via play_audio_file.
Paired with the Ctrl+B keybinding fix in the prior commit, the TUI
voice pipeline now works end-to-end for the first time.
When the user runs /voice and then presses Ctrl+B in the TUI, three
handlers collaborate to consume the chord and none of them dispatch
voice.record:
- isAction() is platform-aware — on macOS it requires Cmd (meta/super),
so Ctrl+B fails the match in useInputHandlers and never triggers
voiceStart/voiceStop.
- TextInput's Ctrl+B pass-through list doesn't include 'b', so the
keystroke falls through to the wordMod backward-word branch on Linux
and to the printable-char insertion branch on macOS — the latter is
exactly what timmie reported ("enters a b into the tui").
- /voice emits "voice: on" with no hint, so the user has no way to
know Ctrl+B is the recording toggle.
Introduces isVoiceToggleKey(key, ch) in lib/platform.ts that matches
raw Ctrl+B on every platform (mirrors tips.py and config.yaml's
voice.record_key default) and additionally accepts Cmd+B on macOS so
existing muscle memory keeps working. Wires it into useInputHandlers,
adds Ctrl+B to TextInput's pass-through list so the global handler
actually receives the chord, and appends "press Ctrl+B to record" to
the /voice on message.
Empirically verified with hermes --tui: Ctrl+B no longer leaks 'b'
into the composer and now dispatches the voice.record RPC (the
downstream ImportError for hermes_cli.voice is a separate upstream
bug — follow-up patch).
Complete the contextvars migration by adding HERMES_SESSION_KEY to the
unified _VAR_MAP in session_context.py. Without this, concurrent gateway
handlers race on os.environ["HERMES_SESSION_KEY"].
- Add _SESSION_KEY ContextVar to _VAR_MAP, set_session_vars(), clear_session_vars()
- Wire session_key through _set_session_env() from SessionContext
- Replace os.getenv fallback in tools/approval.py with get_session_env()
(function-level import to avoid cross-layer coupling)
- Keep os.environ set as CLI/cron fallback
Cherry-picked from PR #7878 by 0xbyt4.
Independent halving of width and height caused aspect ratio distortion
for extreme dimensions (e.g. 8000x200 panoramas). When one axis hit the
64px floor, the other kept shrinking — collapsing the ratio toward 1:1.
Use proportional scaling instead: when either dimension hits the floor,
derive the effective scale factor and apply it to both axes.
Add tests for extreme panorama (8000x200) and tall narrow (200x6000)
images to verify aspect ratio preservation.
LLM responses from browser snapshot extraction and vision analysis
could echo back secrets that appeared on screen or in page content.
Input redaction alone is insufficient — the LLM may reproduce secrets
it read from screenshots (which cannot be text-redacted).
Now redact outputs from:
- _extract_relevant_content (auxiliary LLM response)
- browser_vision (vision LLM response)
- camofox_vision (vision LLM response)
Three exfiltration vectors closed:
1. Browser URL exfil — agent could embed secrets in URL params and
navigate to attacker-controlled server. Now scans URLs for known
API key patterns before navigating (browser_navigate, web_extract).
2. Browser snapshot leak — page displaying env vars or API keys would
send secrets to auxiliary LLM via _extract_relevant_content before
run_agent.py's redaction layer sees the result. Now redacts snapshot
text before the auxiliary call.
3. Camofox annotation leak — accessibility tree text sent to vision
LLM could contain secrets visible on screen. Now redacts annotation
context before the vision call.
10 new tests covering URL blocking, snapshot redaction, and annotation
redaction for both browser and camofox backends.
WSL detection was treated as a hard fail, blocking voice mode even when
audio worked via PulseAudio bridge. Now PULSE_SERVER env var presence
makes WSL a soft notice instead of a blocking warning. Device query
failures in WSL with PULSE_SERVER are also treated as non-blocking.
migrate_model_config() was writing `config["model"] = model_str` which
replaces the entire model dict (default, provider, base_url) with a
bare string. This causes 'str' object has no attribute 'get' errors
throughout Hermes when any code does model_cfg.get("default").
Now preserves the existing model dict and only updates the "default"
key, keeping provider/base_url intact.
auth_type was "***" instead of "api_key" and api_key_env_vars was
("OPEN...",) instead of ("OPENCODE_GO_API_KEY",). This was introduced
in 35d948b6 when a secret redaction tool masked these values during
the Kilo Code provider commit. OpenCode Go provider was completely
broken as a result.
If a tool_calls list contains a None entry (from malformed API response,
compression artifact, or corrupt session replay), convert_messages_to_anthropic
crashes with AttributeError: 'NoneType' object has no attribute 'get'.
Skip None and non-dict entries in the tool_calls iteration. Found via
chaos/fuzz testing with mixed valid/invalid tool_call entries.
Anthropic prompt caching splits input into cache_read_input_tokens,
cache_creation_input_tokens, and non-cached input_tokens. The context
counter only read input_tokens (non-cached portion), showing ~3 tokens
instead of the real ~18K total. Now includes cached portions for
Anthropic native provider only — other providers (OpenAI, OpenRouter,
Codex) already include cached tokens in their prompt_tokens field.
Before: 3/200K | 0%
After: 17.7K/200K | 9%
- 429 rate limit and 529 overloaded were incorrectly treated as
non-retryable client errors, causing immediate failure instead of
exponential backoff retry. Users hitting Anthropic rate limits got
silent failures or no response at all.
- Generic "Sorry, I encountered an unexpected error" now includes
error type, details, and status-specific hints (auth, rate limit,
overloaded).
- Failed agent with final_response=None now surfaces the actual
error message instead of returning an empty response.
Checks the full voice environment and reports what's missing:
- Python packages: discord.py, PyNaCl, davey, STT/TTS providers
- System tools: Opus codec (macOS + Linux paths), ffmpeg
- Environment: bot token, allowed users (resolved to usernames), API keys
- Configuration: STT/TTS provider, voice mode state
- Bot permissions: live Discord API check for Connect, Speak, VAD, etc.
All sensitive values are masked. Gracefully handles missing deps,
invalid tokens, API timeouts, and unreachable Discord API.
Tests cover the actual code paths changed in voice fixes:
_on_packet DAVE passthrough (8 tests):
- Known SSRC + DAVE decrypt success → buffered
- Unknown SSRC + DAVE → skip DAVE, passthrough to Opus
- DAVE "Unencrypted" error → passthrough, not dropped
- DAVE other error → packet dropped
- No DAVE session → direct decode
- Bot's own SSRC → ignored (echo prevention)
- Multiple SSRCs → separate buffers
SSRC auto-mapping (6 tests):
- Single allowed user → auto-mapped
- Multiple allowed users → no auto-map
- No allowlist → sole non-bot member inferred
- Unallowed user → rejected
- Only bot in channel → no map
- Auto-map persists across checks
Buffer lifecycle (4 tests):
- Known SSRC completed utterance
- Short buffer ignored
- Recent audio waits
- Stale unknown buffer discarded
TTS playback (10 tests):
- play_tts calls play_in_voice_channel in VC
- play_tts falls through when not in VC
- play_tts wrong channel no match
- Voice input dedup (runner skips)
- Text + voice_mode combinations
- Error/empty response skipped
- Agent TTS tool dedup
UDP keepalive (2 tests):
- Interval within bounds
- Silence frame actually sent via send_packet
When PyNaCl or davey is not installed, joining a voice channel fails
with a raw exception. Now shows a human-readable message pointing
the user to reinstall with voice support.
Closes#1336
After bot leave/rejoin, Discord doesn't resend SPEAKING events for
users already in the channel. This left SSRC unmapped and all audio
was silently dropped by DAVE decrypt.
Fixes:
- Skip DAVE for unknown SSRCs instead of dropping (passthrough)
- Handle "UnencryptedWhenPassthroughDisabled" DAVE errors gracefully
- Auto-infer user_id from sole allowed member in voice channel
- Pass allowed_user_ids to VoiceReceiver for secure inference
Discord drops the UDP voice route after ~60s of silence - no packets
arrive even when users start speaking again. Send an Opus silence
frame every 15s to keep the UDP session alive.
play_tts was returning success without playing anything when bot was
in a voice channel. Now it calls play_in_voice_channel directly.
Simplified skip_double dedup: base adapter handles voice input TTS
via play_tts (which now works for VC), runner skips to avoid double.
pattern_key was derived by splitting the regex on \b and taking [1],
so patterns starting with the same word (e.g. find -exec rm and
find -delete) produced the same key "find". Approving one silently
approved the other. Using the unique description string as the key
eliminates all collisions.
The fork bomb regex used `()` (empty capture group) and unescaped `{}`
instead of literal `\(\)` and `\{\}`. This meant the classic fork bomb
`:(){ :|:& };:` was never detected. Also added `\s*` between `:` and
`&` and between `;` and trailing `:` to catch whitespace variants.
When the Responses API returns tool call arguments as a dict,
str(dict) produces Python repr with single quotes (e.g. {'key': 'val'})
which is invalid JSON. Downstream json.loads() fails silently and the
tool gets called with empty arguments, losing all parameters.
Affects both function_call and custom_tool_call item types in
_normalize_codex_response().
Moonshot (legacy key) users were shown kimi-for-coding and
kimi-k2-thinking-turbo which only work on the Coding Plan endpoint
(api.kimi.com/coding/v1). Add a separate "moonshot" model list that
excludes plan-specific models.
The rebase added voice prompt checks to _get_tui_prompt_fragments but
the test stub was missing _voice_recording, _voice_processing and
_voice_mode attributes, causing AttributeError.
1. Anthropic + ElevenLabs TTS silence: forward full response to TTS
callback for non-streaming providers (choices first, then native
content blocks fallback).
2. Subprocess timeout kill: play_audio_file now kills the process on
TimeoutExpired instead of leaving zombie processes.
3. Discord disconnect cleanup: leave all voice channels before closing
the client to prevent leaked state.
4. Audio stream leak: close InputStream if stream.start() fails.
5. Race condition: read/write _on_silence_stop under lock in audio
callback thread.
6. _vprint force=True: show API error, retry, and truncation messages
even during streaming TTS.
7. _refresh_level lock: read _voice_recording under _voice_lock.
The mock's app_commands SimpleNamespace lacked choices and Choice attrs,
causing xdist test ordering failures when this mock loaded before
test_discord_slash_commands.
1. Gate _streaming_api_call to chat_completions mode only — Anthropic and
Codex fall back to _interruptible_api_call. Preserve Anthropic base_url
across all client rebuild paths (interrupt, fallback, 401 refresh).
2. Discord VC synthetic events now use chat_type="channel" instead of
defaulting to "dm" — prevents session bleed into DM context.
Authorization runs before echoing transcript. Sanitize @everyone/@here
in voice transcripts.
3. CLI voice prefix ("[Voice input...]") is now API-call-local only —
stripped from returned history so it never persists to session DB or
resumed sessions.
4. /voice off now disables base adapter auto-TTS via _auto_tts_disabled_chats
set — voice input no longer triggers TTS when voice mode is off.
Remove web UI gateway (web.py, tests, docs, toolset, env vars, Platform.WEB
enum) per maintainer request — Nous is building their own official chat UI.
Fix 1: Replace sd.wait() with polling pattern in play_audio_file() to prevent
indefinite hang when audio device stalls (consistent with play_beep()).
Fix 2: Use importlib.util.find_spec() for faster_whisper/openai availability
checks instead of module-level imports that trigger heavy native library
loading (CUDA/cuDNN) at import time.
Fix 3: Remove inspect.signature() hack in _send_voice_reply() — add **kwargs
to Telegram send_voice() so all adapters accept metadata uniformly.
Fix 4: Make session loading resilient to removed platform enum values — skip
entries with unknown platforms instead of crashing the entire gateway.
- web.py: pass stt_model from config like discord.py and run.py do
- run.py: match new error messages (No STT provider / not set)
- _transcribe_local: add missing "provider": "local" to return dict
Move stream close outside the lock in shutdown() to prevent deadlock
when audio callback tries to acquire the same lock. Replace single
t.join(timeout) with a polling loop (0.1s intervals) so KeyboardInterrupt
is not blocked during stream cleanup.
Voice status was hardcoded to check API keys only. Now uses the actual
provider resolution (local/groq/openai) so it correctly shows
"local faster-whisper" when installed instead of "Groq" or "MISSING".
Merge main's faster-whisper (local, free) with our Groq support into a
unified three-provider STT pipeline: local > groq > openai.
Provider priority ensures free options are tried first. Each provider
has its own transcriber function with model auto-correction, env-
overridable endpoints, and proper error handling.
74 tests cover the full provider matrix, fallback chains, model
correction, config loading, validation edge cases, and dispatch.
- Change RTP packet logging from INFO to DEBUG level to reduce noise
(SPEAKING events remain at INFO as they are important lifecycle events)
- Use per-session chat_id (web_{session_id}) instead of shared "web"
to isolate conversation context between simultaneous web users
When bound to 127.0.0.1, only show localhost URL instead of listing
unreachable network interfaces. Add hint about WEB_UI_HOST=0.0.0.0
for phone/tablet access. Add VPN/multi-interface and token exposure
tests (11 new tests).
Only print the access token when auto-generated (user needs it to
log in). When set via WEB_UI_TOKEN env var, just confirm it is set
without exposing the value in console output.
- Path traversal sanitization (Path.name strips ../)
- Media endpoint authentication (401 without token, 404 on traversal)
- hmac.compare_digest usage verification (no == for tokens)
- DOMPurify XSS prevention in HTML template
- Default bind 127.0.0.1 (adapter and config)
- /remote-control token hiding in group chats
- Opus find_library instead of hardcoded paths
- Opus decode error logging (no silent swallow)
- Interrupt _vprint force=True on all 6 calls
- Anthropic interrupt handler in both API call paths
- Update test_web_defaults for new 127.0.0.1 default
- Use hmac.compare_digest for timing-safe token comparison (3 endpoints)
- Default bind to 127.0.0.1 instead of 0.0.0.0
- Sanitize upload filenames with Path.name to prevent path traversal
- Add DOMPurify to sanitize marked.parse() output against XSS
- Replace add_static with authenticated media handler
- Hide token in group chats for /remote-control command
- Use ctypes.util.find_library for Opus instead of hardcoded paths
- Add force=True to 5 interrupt _vprint calls for visibility
- Log Opus decode errors and voice restart failures instead of swallowing
Rebase auto-merge silently overwrote main's Anthropic-aware interrupt
handler with the older OpenAI-only version. Without this fix, interrupting
an Anthropic API call closes the wrong client and leaves token generation
running on the Anthropic side.
Duplicated YAML config parsing for stt.model existed in gateway/run.py
and gateway/platforms/discord.py. Moved to a single helper in
transcription_tools.py and added 5 tests covering all edge cases.
Code fixes:
- STT model, Groq base URL, and OpenAI STT base URL are now
configurable via env vars (STT_GROQ_MODEL, STT_OPENAI_MODEL,
GROQ_BASE_URL, STT_OPENAI_BASE_URL) instead of hardcoded
- Gateway and Discord VC now read stt.model from config.yaml
(previously only CLI did this — gateway always used defaults)
Doc fixes:
- voice-mode.md: move Web UI troubleshooting to web.md (was duplicated)
- voice-mode.md: simplify "How It Works" for end users (remove NaCl,
DAVE, RTP internals)
- voice-mode.md: clarify STT priority (OpenAI used first if both keys
set, Groq recommended for free tier)
- voice-mode.md: document new STT env overrides in config reference
- web.md: remove duplicate Quick Start / Step 1-3 sections
- web.md: add mobile HTTPS mic workarounds (moved from voice-mode.md)
- web.md: clarify STT fallback order
- voice-mode.md: Discord sends native voice bubbles (OGG/Opus flags=8192),
not MP3 file attachments. Falls back to file only if voice API fails.
- discord.md: Bot requires @mention by default in server channels
(DISCORD_REQUIRE_MENTION=true). Previous text incorrectly said no
mention needed.
- index.md: Fix broken ASCII architecture diagram alignment after
adding Web adapter box.
1. VoiceReceiver.stop() now acquires _lock before clearing shared state
to prevent race with _on_packet on the socket reader thread
2. _packet_debug_count moved from class-level to instance-level to avoid
cross-instance race condition in multi-guild setups
3. play_in_voice_channel uses asyncio.get_running_loop() instead of
deprecated asyncio.get_event_loop()
4. _send_voice_reply uses uuid for filenames instead of time-based names
that can collide when two replies happen in the same second
5. Voice timeout now notifies runner via _on_voice_disconnect callback
so runner cleans up _voice_mode state (prevents orphaned TTS replies)
6. play_in_voice_channel adds PLAYBACK_TIMEOUT (120s) to prevent
infinite blocking when FFmpeg callback is never called
7. _send_voice_reply moves temp file cleanup to finally block so files
are always cleaned up even when send_voice/play raises
8. Base adapter auto-TTS wraps play_tts in try/finally with os.remove
to clean up generated audio files after playback
18 new tests (120 total voice tests)
- Add lock protection around VoiceReceiver buffer writes in _on_packet
to prevent race condition with check_silence on different threads
- Wire _voice_input_callback BEFORE join_voice_channel to avoid
losing voice input during the join window
- Add try/except around leave_voice_channel to ensure state cleanup
(voice_mode, callback) even if leave raises an exception
- Guard against empty text after markdown stripping in base.py auto-TTS
- Add 11 tests proving each bug and verifying the fix
- Import from tools.tts_tool instead of reimplementing the logic
- Fix test_truncates_long_text: truncation is the caller's job, not the function's
- Remove unused re import
- macOS firewall may block LAN access to Web UI
- Mobile browsers require HTTPS for microphone API
- Document workarounds: Android Chrome flag, mkcert self-signed cert,
Caddy reverse proxy, SSH tunnel for iOS
Mobile browsers require HTTPS for navigator.mediaDevices API.
Instead of hiding the mic button (confusing UX), show it as dimmed
and display an informative message when tapped explaining the HTTPS
requirement.
When bot is in a Discord voice channel, both base auto-TTS and Discord
play_tts override skip audio. The skip_double guard was also blocking
the runner's _send_voice_reply, resulting in zero audio output in VC.
Now skip_double is overridden when the bot is actively connected to a
voice channel, allowing play_in_voice_channel to handle TTS.
Add comprehensive test matrix covering all platform x input x mode
combinations with full decision table documentation.
- Update TestAutoVoiceReply to include skip_double logic: voice input
is handled by base adapter auto-TTS, gateway runner skips to prevent
duplicate audio
- Add TestDiscordPlayTtsSkip: verifies Discord adapter skips play_tts
when bot is in a voice channel (VC playback handled by runner)
- Add TestWebPlayTts: verifies Web adapter sends invisible play_audio
instead of voice bubble
Base adapter auto-TTS already generates and sends audio for voice
messages in _process_message_background. The gateway runner's
_send_voice_reply was causing double audio on all platforms (not
just Web). Now skip_double applies to any voice input regardless
of platform.
Override play_tts in DiscordAdapter to no-op when connected to a voice
channel for the same guild. The gateway runner already plays TTS audio
in the VC via play_in_voice_channel, so the base adapter's fallback
to send_voice (file attachment) was causing double audio output.
When voice mode is enabled and user sends a voice message on Web UI,
both the base adapter auto-TTS (play_audio) and the gateway voice reply
(send_voice) would fire, causing duplicate audio playback. Skip the
gateway voice reply for Web platform voice input since base adapter
already handles it.
- Document DM vs server channel interaction modes
- Explain @mention requirement and how to select bot user vs role
- Add DISCORD_REQUIRE_MENTION and DISCORD_FREE_RESPONSE_CHANNELS config
- Add troubleshooting entry for bot not responding in server channels
play_tts base class forwards metadata via **kwargs to send_voice,
but Discord and Slack adapters did not accept extra keyword arguments,
causing TypeError and silent message handling failure.
Also fix test_web_defaults to patch correct env var (WEB_UI_TOKEN).
- Voice mode: press mic once to enter, press again to exit
- VAD (Voice Activity Detection) auto-stops recording after 1.5s silence
- Continuous loop: speak → transcribe → agent responds → TTS plays → auto-listen
- Voice mode UI: input bar hides, large mic button centered
- Auto-restart listening when TTS playback finishes
- Fallback: restart listening on text response if no TTS arrives
- Auto-TTS: voice messages get spoken response (audio first, then text)
- STT: Groq Whisper fallback when VOICE_TOOLS_OPENAI_KEY not set
- Futuristic UI: glassmorphism, centered container, purple theme, glow effects
- Voice bubble: custom waveform player with seek and progress
- Invisible TTS playback via play_tts() method (no audio file in chat)
- Add hermes-web toolset with full tool access
- Register Platform.WEB in toolset/config maps
- Update docs for voice conversation feature
Detect all network interfaces instead of relying on UDP trick which
returns VPN IP. Prefers 192.168.x.x/10.x.x.x over VPN ranges.
Shows all available IPs in console output.
Type /remote-control from any platform (Telegram, Discord, etc.) to
instantly start the web UI without restarting the gateway.
- Auto-generates access token if not provided
- Shows URL + token in response
- Optional: /remote-control [port] [token]
- Reports status if already running
- Added to /help command list
New platform adapter that serves a full-featured chat interface via HTTP.
Enables access from any device on the network (phone, tablet, desktop).
Features:
- aiohttp server with WebSocket real-time messaging
- Token-based authentication
- Markdown rendering (marked.js) + code highlighting (highlight.js)
- Voice recording via MediaRecorder API + STT transcription
- Image, voice, and document display
- Typing indicator + message editing (streaming support)
- Mobile responsive dark theme
- Auto-reconnect on disconnect
- Media file cleanup (24h TTL)
Config: WEB_UI_ENABLED=true, WEB_UI_PORT=8765, WEB_UI_TOKEN=<token>
No new dependencies — uses aiohttp already in [messaging] extra.
Cover CLI voice mode, Telegram/Discord auto voice reply, and Discord
voice channel support. Include setup guide with bot permissions, OAuth2
invite URL, privileged intents, system dependencies, and Python packages.
Update discord.md voice messages section with correct STT key reference.
Phase 2 of voice channel support: bot listens to users speaking in VC,
transcribes speech via Groq Whisper, and processes through the agent pipeline.
- Add VoiceReceiver class for RTP packet capture, NaCl/DAVE decryption, Opus decode
- Add silence detection and per-user PCM buffering
- Wire voice input callback from adapter to GatewayRunner
- Fix adapter dict key: use Platform.DISCORD enum instead of string
- Fix guild_id extraction for synthetic voice events via SimpleNamespace raw_message
- Pause/resume receiver during TTS playback to prevent echo
- Send Discord voice messages with flags=8192 and waveform metadata
so they render as native voice bubbles instead of file attachments
- Use .mp3 output path for TTS so edge-tts opus conversion works
correctly (edge always outputs mp3, convert was skipped for .ogg)
- Use actual file_path from TTS result after potential opus conversion
- Register /voice as Discord slash command with mode choices
- Fix _send_voice_reply to handle adapters that don't accept metadata
parameter (Discord) by inspecting the method signature at runtime
- /voice on: reply with voice when user sends voice messages
- /voice tts: reply with voice to all messages
- /voice off: disable, text-only replies
- /voice status: show current mode
- Per-chat state persisted to gateway_voice_mode.json
- Dedup: skips auto-reply if agent already called text_to_speech tool
- drop_pending_updates=True to ignore stale Telegram messages on restart
- 25 tests covering command handler, reply logic, and edge cases
The counter was incremented in start/stop/cancel but never read
anywhere in the codebase. The race condition it was meant to guard
against is practically impossible with the persistent stream design.
- Keep InputStream alive across recordings to avoid CoreAudio hang on
repeated open/close cycles on macOS. New _ensure_stream() creates the
stream once; start()/stop()/cancel() only toggle frame collection.
- Add _close_stream_with_timeout() with daemon thread to prevent
stream.stop()/close() from blocking indefinitely.
- Add generation counter to detect stale stream-open completions after
cancel or restart.
- Run recorder.cancel() in background thread from Ctrl+C handler to
keep the event loop responsive.
- Add shutdown() method called on /voice off to release audio resources.
- Fix silence timer reset during active speech: use dip tolerance for
_resume_start tracker so natural speech pauses (< 0.3s) don't prevent
the silence timer from being reset.
- Update tests to match persistent stream behavior.
AudioRecorder now auto-stops after 15 seconds if no speech is detected
(_has_spoken remains False). In quiet environments where ambient RMS
never exceeds the silence threshold (200), the recording would wait
indefinitely. The new _max_wait parameter fires the silence callback
after the timeout, triggering the normal "No speech detected" flow.
- Set max_retries=0 on the STT OpenAI client. The SDK default (2) honors
Groq's retry-after header (often 53s), blocking the thread for up to
~106s on rate limits. Voice STT should fail fast, not retry silently.
- Stop continuous recording mode after 3 consecutive no-speech cycles to
prevent infinite restart loops when nobody is talking.
0xbyt4