Voice Mode

Real-time voice conversation with your Hermes agent. Tap the mic in chat, speak, and the agent speaks back. No Relay required: on a vanilla connection, speech runs through your Hermes dashboard's audio routes — the same path the official Hermes Desktop voice mode uses — with the server's configured STT/TTS providers.

Two speech routes, picked automatically

• Vanilla Hermes — works on a vanilla Hermes install. The phone talks to your Hermes dashboard; if the dashboard requires sign-in, signing in once under Manage also unlocks voice for that connection.
• Relay — when the optional Relay is paired, voice prefers it: per-profile voice providers, streaming voice output, and the Realtime Agent engine.

You can pin either route under Settings → Voice → Stable STT/TTS Route; the default Auto uses Relay when paired, otherwise Vanilla Hermes.

The stable default engine is Hermes Chat + Voice Output: Hermes owns the chat turn, tools, memory, approvals, and transcript, then the active speech route renders the assistant response to audio. An opt-in Realtime Agent engine is available for experimental provider-native speech work — it requires a paired Relay, is visibly badged as Experimental in Voice Settings, and can be switched off without changing the stable voice behavior.

What It Is

Voice mode is a layer on top of chat. In the stable engine, your voice is transcribed to text, sent through the normal chat flow, and the agent's response is rendered back to speech as it streams in. In the experimental Realtime Agent engine, Android streams mic PCM through the relay to a native realtime provider such as xAI/Grok Voice Agent or OpenAI Realtime, and the relay mirrors the provider's live transcript, audio, and Hermes tool state into the same chat timeline. The MorphingSphere — the same orb you see in the chat empty state — expands to fill the screen and reacts to both your voice and the agent's.

• Your voice drives a subtle blue-purple "listening" state. Gentle breathing, surface wobble with your amplitude.
• The agent's voice drives a dramatic green-teal "speaking" state. The core goes white-hot on peaks, the data ring spins up to 4× speed on loud consonants.

Transcribed messages appear in your chat history as normal messages. Load the session on another device and you'll see the transcript. Realtime Agent mirrors its transcript, Hermes tool state, confirmation prompts, and final response into that same chat timeline so you do not have to exit voice mode to see what happened.

Requirements

On your server:

• Vanilla Hermes route (no Relay): a current hermes-agent whose dashboard exposes the audio endpoints, with stt: and tts: configured in ~/.hermes/config.yaml. If the dashboard is auth-gated, sign in once under Manage on the phone. Older builds without dashboard audio routes show "Not available on this Hermes build" in Voice Settings — update hermes-agent or pair Relay.
• Relay route: Hermes stt: settings for transcription and the relay-managed voice_output: renderer for assistant speech, with the legacy Hermes tts: section as the fallback path. Voice output defaults live in ~/.hermes-relay/config.yaml or in a selected profile's experimental voice_output: section; provider secrets stay server-side.

Common speech output choices:

Provider	API key	Notes
xAI Grok TTS	xAI key or server-side xAI OAuth	First-class relay speech renderer. Built-ins include eve, ara, rex, sal, and leo; relay can also list account custom voices when auth allows it.
OpenAI TTS	VOICE_TOOLS_OPENAI_KEY or OPENAI_API_KEY	Relay speech renderer for OpenAI speech models. Uses OpenAI's documented built-in voice set.
ElevenLabs	ELEVENLABS_API_KEY	Cascaded streaming TTS comparison target. Relay can refresh account voices, models, and languages server-side.
Hermes fallback TTS	Depends on upstream provider	Used if relay voice output fails before audio starts.

Realtime Agent native providers also run relay-side. xai_realtime uses the relay-owned xAI API key or OAuth store. openai_realtime uses OPENAI_REALTIME_API_KEY, OPENAI_API_KEY, or VOICE_TOOLS_OPENAI_KEY on the relay host. These keys are never stored on Android.

Five STT providers are supported:

Provider	API key	Notes
Local (faster-whisper)	None (free)	Default. Runs locally on your server. Models from tiny to large-v3.
Local command	—	Custom whisper binary via HERMES_LOCAL_STT_COMMAND.
Groq	GROQ_API_KEY	Free tier. whisper-large-v3-turbo — very fast.
OpenAI Whisper	VOICE_TOOLS_OPENAI_KEY	whisper-1 or gpt-4o-transcribe.
Mistral Voxtral	MISTRAL_API_KEY	voxtral-mini-latest.

On your phone:

• Microphone permission (requested the first time you tap the mic).
• For Vanilla Hermes voice: a saved dashboard URL and dashboard sign-in when the dashboard requires auth.
• For Relay voice extras or Realtime Agent: a reachable relay at :8767 with the voice routes available.

Relay voice mode can authenticate two ways: a paired Relay session token with voice:config, voice:stt, and voice:tts grants, or the same saved Hermes API key used for API-server fallback chat. That means relay-backed chat+voice-only setups can skip full Relay pairing when only /voice/* is needed. The API-key exception is limited to Relay voice routes and requires HTTPS outside loopback. For a temporary plain-LAN phone test, run hermes relay insecure-api-key on on the relay host, then turn it back off with hermes relay insecure-api-key off.

Before a Relay-backed voice turn is submitted, Android runs a fast relay health preflight. If the host accepts TCP but does not answer HTTP, Relay voice fails quickly with a Connections-facing error instead of sitting in Thinking until the long realtime turn timeout expires. The check is recorded in Settings -> Diagnostics and in the Relay detail sheet's recent activity tail.

Entering Voice Mode

Open a chat and tap the microphone FAB in the bottom-right corner. The first time, Android will ask for microphone permission. Granting it opens the voice mode overlay immediately. Denying it shows a banner at the top — tap Dismiss and you can try again later.

To exit voice mode, tap the X in the top-right of the overlay.

Choose Overlay during an active turn before leaving Hermes to keep Voice available as a wide compact bar over the app you are using. Android will request display-over-other-apps access the first time. Expand it for transcript and response details, minimize it to the existing bubble, or choose Open full voice to return to Hermes without restarting the turn. This optional floating control is separate from Android Assistant mode below.

Interaction Modes

Change in Settings → Voice → Interaction mode. All three share the same overlay — only the mic button's behavior changes.

Tap-to-talk (default)

• Tap the mic to start recording.
• Tap again to stop manually, or stay quiet for ~3 seconds and it'll auto-stop (threshold configurable in Voice Settings).
• Tap while the agent is speaking to interrupt — the current TTS stops and a fresh recording starts. If Barge-in is enabled (see below), you can also just start speaking — no tap needed.

Hold-to-talk

• Press and hold the mic while you speak.
• Release to stop recording and send.
• Good for noisy environments where the silence detector would trip early.

Continuous

• After the agent finishes speaking, the mic automatically re-activates for your next turn.
• A back-and-forth conversation without ever touching the screen.
• Tap the X to exit when you're done.

Streaming Voice Output — Why It Feels Fast

Most voice modes wait for the full response, then synthesize all of it, then play. That's a minimum ~3-second latency on top of the LLM.

Hermes-Relay does balanced client-side voice chunking: as the agent streams its response over SSE, the phone detects complete speech boundaries, batches normal assistant prose into larger natural chunks, and renders those chunks through the relay's /voice/output/* websocket. PCM audio streams directly to Android playback while newer text continues arriving. Short tool/status lines such as "I'm checking that" bypass the batcher so they stay immediate. If streaming voice output fails before audio starts, the app falls back to /voice/synthesize.

Result: first audio starts shortly after the agent begins replying, but normal answers use fewer provider renders so tone, prosody, and volume stay more consistent across the response.

Voice Settings

Settings → Voice is organized into three tabs: Output, Listening, and Advanced. Output keeps the engine, route, provider, model, voice, and save actions together; Listening contains interaction presets and barge-in; Advanced contains STT, host configuration, and the full-engine diagnostic test.

Voice Engine

• Voice engine — Stable Hermes Chat + Voice Output or opt-in Realtime Agent with an Experimental badge.

The active engine controls which provider card appears below. Stable voice shows the Hermes voice-output renderer settings. Realtime voice shows the Realtime Agent provider settings.

Global Voice Controls

These controls always show and apply to both engines:

• Interaction mode — Tap / Hold / Continuous
• Silence threshold — 1-10 seconds, default 3. Only applies in Tap-to-talk mode.
• Auto-TTS — reserved for future: speak all agent responses even when not in voice mode. Currently a placeholder.

Barge-in

Lets you interrupt the agent by speaking while it is thinking or talking. One listener stays active across the whole response, so the microphone does not re-arm between generation and playback. It is on by default to match Hermes Voice and can be disabled at any time.

• Interrupt when I speak — master toggle. Default on.
• Sensitivity — Off / Low / Default / High controls Silero's speech confirmation. Higher values react to quieter or shorter speech; Off disables detection without changing the master switch.
• Room-noise threshold — multiplier applied to the calibrated quiet-room RMS. The upstream default is 3×; higher values require your speech to be louder relative to the room.
• Playback grace — ignores likely echo immediately after playback starts. The upstream default is 0.50 s.
• Resume after interruption — default on. After you interrupt, if you then stay quiet for ~600 ms, the agent resumes reading from the next sentence of its response. A quick breath or "wait, actually…" that you decide not to finish won't throw away its answer. Turn off if you'd rather a hard cut every time.
• VAD diagnostics — optional Logcat detail for calibrated floor, RMS, threshold, phase, and decisions. Leave off unless tuning a device.

Device compatibility. If your phone doesn't support hardware echo cancellation (AcousticEchoCanceler), you'll see a warning badge next to the master toggle: "Your device may have limited echo cancellation. Barge-in quality will vary." You can still enable barge-in, but expect more false triggers from the phone's own speaker feeding back into the mic. Using headphones fixes this entirely — the mic never hears the TTS output, so VAD has nothing to confuse.

Before playback, the app samples roughly 450 ms of room noise and uses its 90th percentile as the floor, with the same bounded generation/playback thresholds, 500 ms grace, and majority decision window as upstream Hermes. Calibration frames do not trigger interruption. The detector follows real playback gaps, uses the quieter generation threshold between output spans, and rearms grace only after a gap of at least one second. The phone's speaker therefore cannot teach calibration the wrong noise floor. As soon as you start speaking, the agent's voice briefly ducks in volume (about 30%). Sustained, model-confirmed speech stops the active generation or playback and captures your replacement request. A false trigger returns to full volume after about 500 ms.

Stop phrases default to exact bare “stop” and can be edited as a comma-separated list; clearing the list disables them. A match ends the active voice chat during generation or playback. Outside voice chat, and for longer requests such as “stop the container,” the words go to the agent normally. Exact pause/resume controls remain specific to Continuous mode.

When you interrupt spoken playback and continue with a new request, Standard Voice privately tells the next model turn that its prior spoken reply was cut off. That one-shot context expires after two minutes and is never added to the visible or persisted transcript. Stopping speech does not cancel a promoted background task; use the task's explicit cancel action or say the explicit background-task cancellation command.

Android Digital Assistant

Hermes can optionally become Android's default Digital Assistant, similar to Home Assistant's assistant integration. This is the supported path for system-mediated background and locked-screen invocation.

1. Open Settings → Voice → Listening.
2. Under Android Digital Assistant, tap Choose Hermes.
3. Grant microphone permission and confirm Hermes in Android's role dialog.
4. Return to Voice settings and enable Background “Hey Hermes” if you want continuous wake detection.

Hermes never changes the default assistant silently. Once selected, the system assistant gesture or power-button shortcut can open a Hermes session even when continuous wake is off. Android presents a real assistant session and hands it to the existing voice flow; Standard voice still uses the upstream Dashboard audio routes and does not require Relay.

Assistant invocation opens as a compact bar over the current screen. Tap the up arrow to expand transcript and response details, then collapse it without ending the turn. Open full voice continues the same request in Hermes' existing Voice screen without restarting the session or handing the microphone to a second recorder. Back collapses an expanded assistant first; Back from the compact bar or Stop closes the assistant turn normally.

The optional background listener uses the same Android-local sherpa model and keeps pre-activation audio on the phone. It releases the microphone before the assistant session starts listening and resumes after the session closes. It is separate from the experimental notification-based listener below; enabling either wake mode disables the other.

To pause continuous listening but keep gesture invocation, turn off Background “Hey Hermes”. To remove Hermes as the default assistant, tap Android settings and choose another assistant or None.

Third-party assistants do not receive Google's dedicated low-power hotword hardware. Continuous local microphone and CPU use can therefore have a noticeable battery cost; leave background wake off if gesture-only invocation is enough. The integration does not use accessibility, overlay permission, full-screen intents, server wake listeners, or Picovoice credentials.

Experimental wake word

Under Voice Settings → Listening, enable Listen for “Hey Hermes” to use the Android-local wake-word preview. It is off by default. The first enable downloads and verifies an English keyword model of about 6 MB. The APK includes the sherpa-onnx runtime but not the model.

Detection runs on the phone: microphone audio is not uploaded before the phrase activates voice. Android requires a user-started microphone foreground service, shows an ongoing privacy notification, and provides a Stop action. The listener does not start at boot. When the phrase is detected, the wake listener releases the microphone before the normal voice flow records anything, stays paused for the voice session, then resumes after voice exits.

The initial preview supports one phrase, “Hey Hermes.” Strictness controls false activations (higher is stricter), Confirmation frames controls how many matching decoder frames are required, and Start a new session chooses between a fresh chat and the selected profile's current session. Profile-specific wake phrases and routing are not supported yet. Continuous local listening uses additional battery and still needs device-specific acoustic testing.

Hermes Chat + Voice Output

Visible when the Hermes Chat + Voice Output engine is selected. It edits streaming voice output defaults for the active profile. Provider, model, voice, language, and sample-rate dropdowns come from the relay's advertised provider metadata. When you pick a provider, the app asks the relay for that provider's latest safe options before saving; account-backed discovery, such as ElevenLabs voice/model lists and paginated xAI custom voices, happens on the server and is cached briefly to avoid repeated provider API calls. OpenAI voice choices are static from the official API docs because OpenAI does not provide a general voice-list endpoint for this surface. The picker groups provider voices, supports search for long voice lists, marks recommended/custom voices where the relay can tell, and validates model/voice/sample-rate compatibility before saving. The provider is summarized with its advertised model and voice counts. The selected model and recommended voices are grouped in one card, with a searchable View all voices picker for larger catalogs. Advanced manual entry lets you type raw IDs for providers or voices the relay does not advertise yet; those save as warnings rather than hard failures unless the relay knows the combination is incompatible. Saving while a profile is selected writes that profile's experimental voice_output: section, so mizuki can use one voice and victor another.

Inline play buttons beside the model and individual voices open an ephemeral voice-output session with the current draft provider/model/voice/language/sample rate. Previewing does not save the draft. Only one preview can play at once; the active row shows the same speaking waveform used by voice mode and its play button becomes Stop. Discard restores the saved profile values, while Save validates and persists the draft independently.

When the streaming provider is xAI Grok TTS, an Expressive speech tags switch appears here. With it on, the relay injects xAI's inline/wrapping tone markers (whisper, laugh, sigh, build/decrease intensity, pitch shifts) into the streamed speech — the streaming-renderer equivalent of the per-request tone tags. It saves per-profile with the other voice-output settings.

The older Hermes tts: config is still shown as fallback provider information and is still used if streaming voice output fails before audio starts.

Enhanced Voice (Gemini & xAI)

When the relay's basic TTS provider supports per-request enhanced control, an Enhanced Voice card appears, titled for the provider (e.g. Enhanced Voice (Gemini)). It lets you steer voice and tone per request without editing server config — the relay maps your choices onto the active provider and never writes them to ~/.hermes/config.yaml. Leave any field on Server default to defer to the saved config.

• Gemini — pick a prebuilt voice, choose the model, and turn on Expressive tone tags (requires a Gemini 3.1 TTS model). With tags on, a hidden rewrite adds cues like [whispers] or [excitedly] to the spoken script; if your server can't run that rewrite it falls back silently to plain speech. An optional Voice direction field takes free-form performance notes ("Warm, calm narrator; unhurried pace.").
• xAI / Grok — type a voice id and turn on Expressive speech tags (xAI's inline/wrapping markers: whisper, laugh, sigh, build/decrease intensity, pitch shifts), with an optional language hint.

Two scoping notes:

• Relay-only. On the Vanilla Hermes (no-plugin) path the dashboard /api/audio/speak endpoint accepts only the text to speak, so enhanced voice there is whatever the server's tts.<provider>.* config sets — change it under Manage, and it applies to every voice turn, not per utterance.
• Which path it drives. This card's per-request overrides ride /voice/synthesize, used when the streaming voice-output renderer above is off or unavailable (and as the fallback when streaming fails). For the streaming renderer — the default for relay playback — xAI's expressive speech tags are controlled by the Expressive speech tags switch in the Hermes Chat + Voice Output card above (saved per-profile via voice_output:), and voice is the voice-output provider voice. Gemini is synthesize-only (the streaming renderer has no Gemini provider). Providers without an enhanced surface (ElevenLabs, MiniMax, Mistral, OpenAI, local engines) stay config-only.

Stable streaming playback can resume during short connection changes. If the phone moves from Wi-Fi to cellular or from LAN to Tailscale while audio is in flight, Android reopens the same relay voice-output session through the current route and the relay replays missed PCM chunks instead of forcing a new render.

Speech-to-Text

Read-only display of the STT provider and model reported by /voice/config. STT still follows the upstream Hermes stt: configuration for the stable Hermes Chat + Voice Output engine, resolved through the active profile where Hermes has one. Realtime Agent uses the selected provider's native realtime transcription instead and does not upload the turn through /voice/transcribe.

Realtime Agent

Visible when the Realtime Agent voice engine is selected. This experimental section saves profile-scoped realtime provider defaults (realtime_voice:) for provider-native testing. Native Realtime Agent providers currently include xai_realtime with grok-voice-latest and openai_realtime with gpt-realtime-2. Victor's profile can select leo for xAI, while OpenAI profiles commonly use voices such as marin or cedar. Provider dropdowns use the same relay-owned option refresh, search, grouping, and validation pattern as Voice Output, plus a native-agent capability flag so lab/render-only providers do not appear as full Realtime Agent choices.

Realtime Agent is still Hermes-first. The provider does not get raw Android bridge tools, direct Hermes tool execution, or ownership of memory. The relay broker exposes only a small tool surface:

• hermes_run_task
• hermes_get_status
• hermes_cancel
• hermes_confirm

When the phone is paired, its Relay session token only authenticates it to the realtime route. Hermes tool calls from the realtime broker use the relay server's local Hermes API credential from config.yaml, .env, or API_SERVER_KEY; the provider never receives the phone's saved API key.

During Hermes tool work, the app shows clean status from the broker instead of dumping raw tool output into voice. Normal timeline rows show the active Hermes tool, completion state, and provider-spoken provenance. Android keeps those status updates UI-only during provider-native realtime playback so a second voice-output stream cannot compete with the realtime provider. The final spoken answer is generated by the realtime provider after Hermes returns a compact result, so tool output is summarized naturally instead of read aloud.

Background tasks

Some requests take a while — research, multi-step work, a long command. Instead of freezing the conversation until they finish, Realtime Agent promotes a slow run to the background: the agent says a short "I'm on it" and you can keep talking, ask something else, or just wait. When the task finishes, the agent speaks the answer. Asking for something explicitly long starts a background task right away.

You control this under Voice Settings → Realtime Agent → Background tasks:

• Promote long tasks — turn the behavior on or off. With it off, the agent waits silently until the task finishes (the old behavior).
• Spoken handoff — whether the agent says a short acknowledgement when a task moves to the background, or just shows it on screen.
• When the answer is ready — Speak it as soon as you're not mid-sentence, Notify and speak when you re-engage, or Show only (no spoken answer).

While a background task is running, a small "working on it" chip stays visible in the voice screen. You can cancel a background task at any time the same way you cancel any turn. Hermes still owns the task end-to-end — promotion only changes when the answer is spoken, never who runs the tools.

Provider-native Android paths stream mic PCM to a relay-owned realtime provider WebSocket session. Android commits the captured utterance, the active provider owns input transcription and speech generation, and Hermes still owns profile binding, session history, memory, tools, confirmation prompts, cancellation policy, and transcript persistence. The relay sends the provider only the approved Hermes function schemas, returns compact tool results, and waits for Android playback to drain before asking the provider for post-tool narration. If the provider disconnects or quality is poor, switch Voice engine back to Hermes Chat + Voice Output; existing voice routes and settings remain intact.

Realtime Agent sessions use the same resume window during short connection changes. If the phone moves from Wi-Fi to cellular or from LAN to Tailscale while a turn is thinking or speaking, Android follows the app's current relay route signal and can reopen the same relay voice session before the old websocket times out. It sends a session resume token and its last received audio/event IDs, and the relay replays missed status or PCM chunks instead of starting a second Hermes run. Android waits for the relay's resume confirmation before sending another recorded turn, and retries for a bounded window that starts when the route is lost. If that window expires, the app detaches the old task status, shows a recoverable voice error, and the next tap starts a fresh turn; a durable Hermes task may still report through chat or notification.

Voice turns include interface context. In stable voice mode the chat agent is told the turn came through Hermes Chat + Voice Output; in Realtime Agent mode the provider and Hermes broker are told the active path is realtime_agent, including provider, model, voice, profile, and relay-local date/time. If you ask which path is active or what today's date is, the agent should answer from that context.

Realtime Agent also receives recent shared chat context from the current timeline. That means switching from normal chat voice to Realtime Agent, then back again, should feel like one conversation. If the realtime provider answers a simple prompt-contained question directly, the app syncs that local provider-only turn into the next Hermes chat turn. If the turn used Hermes tools, Hermes already has the canonical record and the app does not duplicate the tool output.

For research, news, current facts beyond the injected date/time, live checks, or anything else the realtime provider cannot know from the active context, the provider is instructed to call Hermes instead of guessing from model knowledge. That also covers latest/versioned data, device or desktop state, personal or project context, side effects, precision-sensitive answers, explicit check/verify/look-up requests, media or artifact handling, and follow-up references like "this", "that", "it", or "that integration" that need durable chat/session context. Hermes performs the governed check, then the provider speaks a concise summary from Hermes' returned answer.

Realtime Agent also asks the provider to format speech for listening: dates, times, currency, percentages, versions, measurements, counts, paths, URLs, IDs, JSON, logs, tables, and dense numeric strings should be summarized naturally instead of read character by character. When raw values are not useful aloud, the voice can say something like "plus a few IDs and raw values" and keep the meaning front and center.

The default Realtime Agent timeline stays user-facing: live transcript, assistant speech, path badges, confirmation state, and compact tool rows. Turn on Detailed trace in Voice settings to show compact Hermes status and result provenance for debugging. Full raw traces stay in the relay run logs.

Reliable, low-latency playback. Realtime audio plays from the first frame with minimal latency. Earlier builds could drop or stutter the first turn because the AudioTrack deep buffer cold-started with its playback head parked at zero; the streaming buffer was reduced to a low-latency size, the prebuffer threshold retuned, and the preroll force-start removed. Playback now records a time-to-first-audio metric, logs the requested-vs-actual buffer size, runs a first-frame watchdog, and cross-checks drain drift. When something goes wrong, those signals land in Settings -> Diagnostics instead of presenting as silent dead air.

Test Current Engine

When Hermes Chat + Voice Output is selected, plays a sample sentence ("Hello, this is Hermes. Voice mode is working.") through the currently saved profile voice-output path. When Realtime Agent is selected, opens a provider-native /voice/realtime-agent/* session and plays a short realtime sample through the relay. The fallback TTS card is global and remains visible for both engines as the stable speech safety net.

Voice Lab

The realtime voice test screen (the "Voice Lab") is a developer-facing harness for exercising the realtime path in isolation, with two clearly separated demos:

• Text demo — plays raw provider TTS directly, with no agent in the loop. Use it to confirm the provider, voice, and streaming playback are healthy.
• Mic demo — runs the full agent path: real speech recognition, Hermes brokering, and a spoken reply. Capture is tap-to-record / tap-to-stop, so you control exactly when the utterance is committed.

The Voice Lab waveform follows the playback cursor — it is driven by the player's amplitude at the current playback position rather than by socket-arrival time, so the visual matches what you actually hear. A scripts/realtime-voice-lab-smoke.ps1 script automates a quick end-to-end check of the lab routes.

Troubleshooting

"Microphone permission is required for voice mode" — tap the mic FAB again. If Android doesn't show a permission prompt (you denied twice), go to Android Settings → Apps → Hermes-Relay → Permissions → Microphone and enable it.

No audio plays when the agent speaks — check Settings → Voice → Hermes Chat + Voice Output. The Render path row tells you which path renders speech: streaming (/voice/output) when the renderer is active, or basic synthesize (/voice/synthesize) when it's off or unavailable (the streaming "Expressive speech tags" toggle only applies on the streaming path). If the provider shows "Unknown" or the profile scope is not what you expected, the relay's /voice/output/config or /voice/config endpoint returned an error or fell back. Try Test Current Engine; the error message will tell you what's wrong. Settings → Diagnostics also logs the active render path per session.

Transcript comes back empty or garbled — check Settings → Voice → Speech-to-Text. Same diagnostic: Test Current Engine won't catch stable STT issues, but you can verify by speaking a slow, clear phrase and watching the transcribed text appear in the overlay. If it's wrong, your STT model or language setting is off.

The sphere doesn't react during Speaking on my device — some Android OEMs refuse to construct the Visualizer audio-effect object even with MODIFY_AUDIO_SETTINGS granted. Hermes-Relay catches this and falls back to a flat amplitude — voice mode still works, the orb just won't pulse with the agent's voice. Listening-mode amplitude (from your mic) is unaffected.

"Relay returned 503" — your server doesn't have the provider's optional dependencies installed. SSH into the server and pip install whichever provider you configured (e.g. pip install edge-tts for Edge, pip install elevenlabs for ElevenLabs, or pip install faster-whisper for local STT).

"Relay returned 413" on synthesize — you're trying to synthesize more than 5000 characters at once. This is a safety cap on the relay side to avoid runaway TTS costs. Client-side sentence chunking should normally keep individual requests well under this, so a 413 usually means the agent returned one enormous uninterrupted sentence.

"That pairing code was already used" — Relay pairing codes are one-shot. Generate a fresh QR from the dashboard Relay tab or hermes pair and scan again. If you only need chat plus standard voice, skip Relay pairing and use the Dashboard/Gateway connection; the same dashboard session authorizes both.

Privacy Note

Voice audio goes through the active speech route: the upstream Dashboard for standard Vanilla Hermes voice, or Relay for optional enhanced/realtime engines. From there it reaches whichever provider you configured. If you're using a cloud provider (ElevenLabs, OpenAI, Groq, Mistral), your audio goes to them. If you're using local providers (faster-whisper, NeuTTS, Edge TTS), nothing leaves your network.

The mp3 files returned from /voice/synthesize are cached briefly in the app's cache directory and cleared automatically as new ones arrive (capped at 6 at a time). On the server side, the relay writes each /voice/synthesize render to a private temp file and deletes it after streaming, so relay synthesis no longer accumulates files in ~/voice-memos/ (other Hermes agent voice features may still use that directory).