Echo

Between breath and hearing, a voice can vanish.

A spoken answer can be generated yet never heard. Echo follows an utterance from deliberately armed capture to transcription, an ordinary Agent step, optional synthesis, and observed delivery. Source, consent, language, interruption, and playback chronology remain attached throughout.

The first passage is visible push-to-talk and record-and-send. Audio admission is Partial, but no Echo speech path runs; State owns delivery and Audio owns the accepted law.

The resonance path

The first profile, record-and-send, carries one utterance through these steps:

visible bounded capture → immutable audio artifact → eligible Ear or audio-capable Mind → attributed transcript or native observation → ordinary Agent step → optional Voice → delivery evidence

Capture must be visible, bounded, and consented. Large bytes remain outside Graph checkpoints and queues.

Three distinctions keep the path honest:

• ArtifactRef carries metadata and grants no access. The explicit local custody seam can retain supplied WAV bytes and feed Echo's inspector after separate authorization and admission; it captures no recording and supplies no playback or speech path.
• Reliquary's accepted design preserves source and derivation; the first local profile retains source bytes and attributed claims, without a derivative graph. A transcript is an attributed interpretation, not a replacement for its recording.
• Synthesized audio is a new artifact. Generation proves neither delivery nor playback; receipts must state what actually reached the listener.

Later half- and full-duplex profiles must preserve one monotonic speech timeline through barge-in, contested turn ownership, interruption, cancellation, and reconnect. A live socket cannot supply those semantics.

Invalid audio, insufficient authority, and policy-ineligible remote egress are refused. Only an otherwise eligible managed provider that is not warm may cause the requesting Run to enter ordinary Graph Stasis while Orchestrator drains and readies the affected Animators. Echo cannot silently fall back to a remote service, infer permanent microphone authority, inflate priority, or revoke another grant or lease.

The organs of resonance

An Echo deployment composes five roles independently. Audio is material, never a capability family. The labels stt (speech-to-text), tts (text-to-speech), and chat below belong to current v1 compatibility; the Designed general-service interfaces define their own request and return contracts.

Echo's boundary is speech: utterance capture, STT, spoken-output synthesis, delivery, and playback chronology. It does not own music, sung performance, ambience, sound effects, picture/world sound, or generic audio understanding or generation. A model or engine that natively emits audio, video, or both gains no broader Echo authority; every non-speech facet remains with its application and technical effect owner.

Applications may use speech without owning Echo

Echo owns speech contracts and lifecycle facts, not every work containing a speaking voice. Riffmaw owns sung and otherwise musical performance; Language Edition owns casting, linguistic adaptation, timing, captions, and acceptance for one timed-language edition; Avatar may reference an exact eligible acoustic voice inside a presentation profile; and Broadcast owns canonical source words, editorial relation, and publication. Each application references exact Echo attempts and artifacts rather than copying its capture, synthesis, delivery, or playback chronology.

The Companion device boundary

A future Companion client may own device capture and playback controls as Echo's physical speech endpoint. A private Tether supplies reachability, not caller authority: Ward still authenticates and authorizes, while capture and consent remain visible. A remote Portal additionally requires eligible classification, egress policy, consent where required, and bounded cost. These are laws of the accepted design, not claims of delivered behavior.

First engine profile

The first planned local profile deliberately admits one inference engine:

This is designed direction, not delivered integration. The initial implementation should prove speech-audio-to-text and text-to-speech-audio before exposing audio.cpp speech-voice conversion, cloning, alignment, VAD, diarization, codecs, or pipelines. Those functions become separately declared tool surfaces only after their model, inputs, outputs, custody, authority, and cancellation behavior are baked; the presence of a family in an upstream catalogue grants no LychD capability.

One connector may serve several compatible model profiles, but every profile keeps its own model identity, language facts, license, backend, resource measurements, and readiness. A language can be declared from upstream metadata without being verified by LychD. Routing uses an explicit request first, then ordered Principal or Persona preferences, then admitted detection; missing output-language support refuses rather than silently translating.

The first bake

The smallest honest evaluation fixes one runtime revision and model digest, then records:

• cold start, readiness, clean shutdown, cancellation, CPU/GPU placement, RAM/VRAM peak, and isolation from a resident Mind;
• transcription accuracy, endpoint truncation, names, numbers, mixed-language speech, confidence, and provenance for each language claimed as verified;
• synthesis latency, first audio, intelligibility, pronunciation, numbers, dates, currency, addresses, text normalization, and playback completion for each output locale and voice;
• exact OpenAI-compatible fields and errors rather than compatibility by endpoint name; and
• license and consent receipts before any voice-reference or cloning experiment.

Slovak is the first operator-selected reference bake, not a Core hardcode. Other users may order different input and output preferences, but every routed language owes the same evidence.

Watched alternatives

The first profile does not implement these paths:

NeMo-Speech.cpp may later reuse the batch connector through an explicit dialect, while its live WebSocket still needs a Resonance Session adapter. Whisper, Piper, and openWakeWord are not installed merely as speculative redundancy. No watched candidate is an automatic fallback, and no failed local attempt silently changes engine, language, or egress.