# ishaqahmed02/voice-agent-chat architecture

> Architecture diagram of the GitHub repository ishaqahmed02/voice-agent-chat, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2025-09-15). It is an AI-made overview: check details against the source.

- Interactive diagram: https://gitdiagram.com/ishaqahmed02/voice-agent-chat
- Repository: https://github.com/ishaqahmed02/voice-agent-chat

## Overview

##### Project type and purpose
- Key clues:
- sts.py – likely the “speech-to-speech” or “speech-to-system” orchestrator
- whisper/ – a local module (whisper) containing audio processing scripts (stt.py, live.py, dlive.py) and generated artifacts (.json, .txt, .srt, .tsv, .vtt, .wav)
- output.wav – suggests the tool generates synthesized or transformed audio

##### File structure
- Top level:
- README.md
- sts.py (main script)
- output.wav (resulting audio)
- whisper/ directory:
- stt.py – speech‐to‐text component
- live.py – live (streaming) STT wrapper
- dlive.py – possibly “dual” or “distributed” live mode
- audio.* – transcripts in various formats
- audio.wav – intermediate or input WAV

- No separate frontend/backend folders – it’s a single‐process or small service tool.

##### README for additional insights
- README only gives the project name, so the architecture must be inferred from filenames and typical Whisper usage.
- Expect modules for: audio capture, transcription (STT), dialogue manager (chat agent), and audio output (TTS).

## Diagram (Mermaid)

```mermaid
flowchart TD
    %% Audio input
    AudioSource["Audio Source\n(microphone or input WAV)"]:::processing

    %% Whisper STT modules
    subgraph "Whisper STT Modules"
        direction TB
        Stt["Batch STT\n(stt.py)"]:::processing
        Live["Live STT\n(live.py)"]:::processing
        Dlive["Dual-Live STT\n(dlive.py)"]:::processing
    end

    %% Transcript artifacts
    subgraph "Transcript Artifacts"
        direction TB
        AudioJson["audio.json"]:::datastore
        AudioTxt["audio.txt"]:::datastore
        AudioSrt["audio.srt"]:::datastore
        AudioTsv["audio.tsv"]:::datastore
        AudioVtt["audio.vtt"]:::datastore
    end

    %% Orchestration and Dialogue
    Orchestrator["Orchestration Layer\nsts.py"]:::orchestrator

    %% TTS and final output
    TTS["TTS Module"]:::processing
    FinalAudio["Final Audio Output\n(output.wav)"]:::datastore

    %% Flow connections
    AudioSource -->|".wav"| Stt
    AudioSource -->|".wav (stream)"| Live
    AudioSource -->|".wav (dual-stream)"| Dlive

    Stt -->|"transcript files"| AudioJson
    Stt -->|"transcript files"| AudioTxt
    Stt -->|"transcript files"| AudioSrt
    Stt -->|"transcript files"| AudioTsv
    Stt -->|"transcript files"| AudioVtt

    Live -->|"transcript files"| AudioJson
    Live -->|"transcript files"| AudioTxt
    Live -->|"transcript files"| AudioSrt
    Live -->|"transcript files"| AudioTsv
    Live -->|"transcript files"| AudioVtt

    Dlive -->|"transcript files"| AudioJson
    Dlive -->|"transcript files"| AudioTxt
    Dlive -->|"transcript files"| AudioSrt
    Dlive -->|"transcript files"| AudioTsv
    Dlive -->|"transcript files"| AudioVtt

    AudioJson -->|"transcripts"| Orchestrator
    AudioTxt -->|"transcripts"| Orchestrator
    AudioSrt -->|"transcripts"| Orchestrator
    AudioTsv -->|"transcripts"| Orchestrator
    AudioVtt -->|"transcripts"| Orchestrator

    Orchestrator -->|"text response"| TTS
    TTS -->|".wav"| FinalAudio

    %% Click events
    click Orchestrator "https://github.com/ishaqahmed02/voice-agent-chat/blob/main/sts.py"
    click Stt "https://github.com/ishaqahmed02/voice-agent-chat/blob/main/whisper/stt.py"
    click Live "https://github.com/ishaqahmed02/voice-agent-chat/blob/main/whisper/live.py"
    click Dlive "https://github.com/ishaqahmed02/voice-agent-chat/blob/main/whisper/dlive.py"
    click AudioJson "https://github.com/ishaqahmed02/voice-agent-chat/blob/main/whisper/audio.json"
    click AudioTxt "https://github.com/ishaqahmed02/voice-agent-chat/blob/main/whisper/audio.txt"
    click AudioSrt "https://github.com/ishaqahmed02/voice-agent-chat/blob/main/whisper/audio.srt"
    click AudioTsv "https://github.com/ishaqahmed02/voice-agent-chat/blob/main/whisper/audio.tsv"
    click AudioVtt "https://github.com/ishaqahmed02/voice-agent-chat/blob/main/whisper/audio.vtt"
    click AudioSource "https://github.com/ishaqahmed02/voice-agent-chat/blob/main/whisper/audio.wav"
    click FinalAudio "https://github.com/ishaqahmed02/voice-agent-chat/blob/main/output.wav"

    %% Styles
    classDef processing fill:#b6fcb6,stroke:#2d7f2d,color:#000
    classDef datastore fill:#cce5ff,stroke:#1a5276,color:#000
    classDef orchestrator fill:#ffedcc,stroke:#975f0f,color:#000
```
