PODPIPELINE

2 Squirrels AI

Technology Stack

End-to-end AI-powered podcast automation platform with multi-agent orchestration, Human-in-the-Loop (HITL) approval checkpoints, intelligent content discovery, and automated distribution to major platforms.

Orchestration LangGraph
LLM Claude / OpenAI
Web Search Tavily API
TTS Chatterbox / Kokoro
Backend NestJS
Queue BullMQ

Workflow 1: End-to-End Podcast Production Pipeline

%%{init: {'theme': 'dark', 'themeVariables': { 'primaryColor': '#C17852', 'primaryTextColor': '#F0F6FC', 'primaryBorderColor': '#4A5E32', 'lineColor': '#E6C98F', 'secondaryColor': '#161B22', 'tertiaryColor': '#0D1117', 'background': '#0D1117', 'mainBkg': '#161B22', 'nodeBorder': '#4A5E32', 'clusterBkg': '#161B22', 'clusterBorder': '#4A5E32', 'titleColor': '#E6C98F', 'edgeLabelBackground': '#161B22'}}}%%
flowchart TB
    subgraph Input["📥 USER INPUT"]
        SUB["🎯 Subjects + Params"]
    end

    subgraph Discovery["🔍 CONTENT DISCOVERY"]
        SEARCH["🌐 Web Search
Tavily API"]
        RANK["📊 Article Ranking
& Selection"]
    end

    subgraph HITL1["👤 HITL CHECKPOINT 1"]
        APP1{"✅ Human
Approval?"}
    end

    subgraph Processing["⚙️ CONTENT PROCESSING"]
        SUM["📝 Content
Summarization"]
        HOST["🎭 Host Personality
Generation"]
        SCRIPT["📜 Script
Generation"]
    end

    subgraph HITL2["👤 HITL CHECKPOINT 2"]
        APP2{"✅ Script
Review?"}
    end

    subgraph Production["🎵 AUDIO PRODUCTION"]
        TTS["🎤 TTS Audio
Generation"]
        POST["🎚️ Audio
Post-Processing"]
    end

    subgraph HITL3["👤 HITL CHECKPOINT 3"]
        APP3{"✅ Final
Approval?"}
    end

    subgraph Distribution["📤 DISTRIBUTION"]
        PUBLISH["🚀 Platform
Distribution"]
        ANALYTICS["📈 Analytics &
Monitoring"]
    end

    SUB --> SEARCH
    SEARCH --> RANK
    RANK --> APP1
    APP1 -->|"Approved"| SUM
    APP1 -->|"Rejected"| SEARCH
    SUM --> HOST
    HOST --> SCRIPT
    SCRIPT --> APP2
    APP2 -->|"Approved"| TTS
    APP2 -->|"Edit"| SCRIPT
    TTS --> POST
    POST --> APP3
    APP3 -->|"Approved"| PUBLISH
    APP3 -->|"Reject"| POST
    PUBLISH --> ANALYTICS

HITL Checkpoints: Three human approval gates ensure quality at critical junctures - content selection, script review, and final audio approval before publishing.

Workflow 2: LangGraph Content Discovery State Machine

%%{init: {'theme': 'dark', 'themeVariables': { 'primaryColor': '#C17852', 'primaryTextColor': '#F0F6FC', 'primaryBorderColor': '#4A5E32', 'lineColor': '#E6C98F', 'secondaryColor': '#161B22', 'tertiaryColor': '#0D1117', 'background': '#0D1117', 'mainBkg': '#161B22', 'nodeBorder': '#4A5E32', 'clusterBkg': '#161B22', 'clusterBorder': '#4A5E32', 'titleColor': '#E6C98F', 'edgeLabelBackground': '#161B22'}}}%%
flowchart TB
    subgraph State["📊 DISCOVERY STATE"]
        direction LR
        S1["projectId"]
        S2["parameters"]
        S3["searchResults"]
        S4["articles"]
        S5["rankedArticles"]
        S6["approvalStatus"]
    end

    subgraph Agents["🤖 LANGGRAPH AGENTS"]
        direction TB
        AG1["🔎 searchAgent
Web Search"]
        AG2["📄 extractAgent
Content Extraction"]
        AG3["⭐ rankAgent
Ensemble Scoring"]
    end

    subgraph HITL["👤 HUMAN IN THE LOOP"]
        WAIT["⏳ wait_for_approval"]
        REFINE["🔄 refine_search
Based on Feedback"]
    end

    subgraph Edges["🔀 CONDITIONAL EDGES"]
        COND{"approvalStatus?"}
    end

    subgraph Output["📤 OUTPUT"]
        END1(("🏁 END"))
    end

    State --> AG1
    AG1 -->|"addEdge"| AG2
    AG2 -->|"addEdge"| AG3
    AG3 -->|"addEdge"| WAIT
    WAIT --> COND
    COND -->|"approved"| END1
    COND -->|"rejected"| REFINE
    COND -->|"pending"| WAIT
    REFINE -->|"addEdge"| AG1

Workflow 3: AI-Powered Article Ranking System

%%{init: {'theme': 'dark', 'themeVariables': { 'primaryColor': '#C17852', 'primaryTextColor': '#F0F6FC', 'primaryBorderColor': '#4A5E32', 'lineColor': '#E6C98F', 'secondaryColor': '#161B22', 'tertiaryColor': '#0D1117', 'background': '#0D1117', 'mainBkg': '#161B22', 'nodeBorder': '#4A5E32', 'clusterBkg': '#161B22', 'clusterBorder': '#4A5E32', 'titleColor': '#E6C98F', 'edgeLabelBackground': '#161B22'}}}%%
flowchart LR
    subgraph Input["📥 ARTICLES"]
        ART["📄 Raw Articles"]
    end

    subgraph Scoring["📊 MULTI-DIMENSIONAL SCORING"]
        direction TB
        SC1["🎯 Relevance
0-100"]
        SC2["⏰ Freshness
0-100"]
        SC3["🏛️ Authority
0-100"]
        SC4["💡 Engagement
0-100"]
        SC5["✨ Uniqueness
0-100"]
        SC6["🔥 Controversy
0-100"]
        SC7["❤️ Emotional
0-100"]
    end

    subgraph Ensemble["🧠 ENSEMBLE AGENT"]
        WEIGHT["⚖️ Weighted
Composite"]
        CONF["📈 Confidence
Score"]
    end

    subgraph Output["📤 RANKED OUTPUT"]
        RANK["🏆 Ranked Articles
with Scores"]
    end

    ART --> SC1 & SC2 & SC3 & SC4 & SC5 & SC6 & SC7
    SC1 & SC2 & SC3 & SC4 & SC5 & SC6 & SC7 --> WEIGHT
    WEIGHT --> CONF
    CONF --> RANK

Workflow 4: AI Script Generation Pipeline

%%{init: {'theme': 'dark', 'themeVariables': { 'primaryColor': '#C17852', 'primaryTextColor': '#F0F6FC', 'primaryBorderColor': '#4A5E32', 'lineColor': '#E6C98F', 'secondaryColor': '#161B22', 'tertiaryColor': '#0D1117', 'background': '#0D1117', 'mainBkg': '#161B22', 'nodeBorder': '#4A5E32', 'clusterBkg': '#161B22', 'clusterBorder': '#4A5E32', 'titleColor': '#E6C98F', 'edgeLabelBackground': '#161B22'}}}%%
flowchart TB
    subgraph Input["📥 INPUTS"]
        ARTICLES["📄 Approved
Articles"]
        FORMAT["🎬 Show
Format"]
    end

    subgraph Summarization["📝 SUMMARIZATION"]
        SUM1["📰 Article Summary
Agent"]
        SUM2["🔗 Multi-Doc
Synthesis Agent"]
    end

    subgraph Personality["🎭 PERSONALITY"]
        direction TB
        TRAITS["📊 Personality Matrix
Intelligence, Tone, Energy"]
        VOCAB["📖 Vocabulary
& Speech Patterns"]
        QUIRKS["🎪 Catchphrases
& Quirks"]
    end

    subgraph Generation["✍️ SCRIPT GENERATION"]
        STRUCT["🏗️ Structure Agent
Segments & Pacing"]
        DIALOG["💬 Dialogue Agent
Host Conversations"]
        QA["✅ Quality Check
Agent"]
    end

    subgraph Metrics["📊 SCRIPT METRICS"]
        M1["📏 Word Count"]
        M2["⏱️ Duration Est."]
        M3["😊 Sentiment"]
        M4["📈 Energy Profile"]
    end

    ARTICLES --> SUM1
    SUM1 --> SUM2
    FORMAT --> STRUCT
    TRAITS & VOCAB & QUIRKS --> DIALOG
    SUM2 --> STRUCT
    STRUCT --> DIALOG
    DIALOG --> QA
    QA --> M1 & M2 & M3 & M4

Workflow 5: TTS Audio Production Pipeline

%%{init: {'theme': 'dark', 'themeVariables': { 'primaryColor': '#C17852', 'primaryTextColor': '#F0F6FC', 'primaryBorderColor': '#4A5E32', 'lineColor': '#E6C98F', 'secondaryColor': '#161B22', 'tertiaryColor': '#0D1117', 'background': '#0D1117', 'mainBkg': '#161B22', 'nodeBorder': '#4A5E32', 'clusterBkg': '#161B22', 'clusterBorder': '#4A5E32', 'titleColor': '#E6C98F', 'edgeLabelBackground': '#161B22'}}}%%
flowchart LR
    subgraph Script["📜 APPROVED SCRIPT"]
        SEG["📝 Script
Segments"]
    end

    subgraph Config["⚙️ TTS CONFIG"]
        ENG["🔧 Engine Selection
Chatterbox / Kokoro"]
        VOICE["🎭 Voice Mapping
Host → Voice ID"]
        PARAMS["🎛️ Parameters
Pitch, Speed, Emotion"]
    end

    subgraph Generation["🎤 TTS GENERATION"]
        TTS1["🔊 Generate
Audio Segments"]
        QUEUE["📋 BullMQ
Job Queue"]
    end

    subgraph PostProcess["🎚️ POST-PROCESSING"]
        NORM["📊 Volume
Normalization"]
        NOISE["🔇 Noise
Suppression"]
        COMPRESS["📦 Compression"]
    end

    subgraph Assembly["🎵 ASSEMBLY"]
        INTRO["🎼 Intro
Music"]
        CONCAT["🔗 Concatenate
Segments"]
        OUTRO["🎼 Outro
Music"]
        FADE["📉 Fade
In/Out"]
    end

    subgraph Output["📤 FINAL"]
        MP3["🎧 Episode.mp3"]
        META["📋 Metadata
ID3 Tags"]
    end

    SEG --> ENG
    ENG --> VOICE
    VOICE --> PARAMS
    PARAMS --> TTS1
    TTS1 --> QUEUE
    QUEUE --> NORM
    NORM --> NOISE
    NOISE --> COMPRESS
    COMPRESS --> CONCAT
    INTRO --> CONCAT
    OUTRO --> CONCAT
    CONCAT --> FADE
    FADE --> MP3
    FADE --> META

Workflow 6: Multi-Platform Distribution System

%%{init: {'theme': 'dark', 'themeVariables': { 'primaryColor': '#C17852', 'primaryTextColor': '#F0F6FC', 'primaryBorderColor': '#4A5E32', 'lineColor': '#E6C98F', 'secondaryColor': '#161B22', 'tertiaryColor': '#0D1117', 'background': '#0D1117', 'mainBkg': '#161B22', 'nodeBorder': '#4A5E32', 'clusterBkg': '#161B22', 'clusterBorder': '#4A5E32', 'titleColor': '#E6C98F', 'edgeLabelBackground': '#161B22'}}}%%
flowchart TB
    subgraph Episode["🎧 FINAL EPISODE"]
        AUDIO["🔊 Audio File"]
        META["📋 Metadata"]
        ART["🖼️ Cover Art"]
    end

    subgraph MetadataAgent["🤖 METADATA AGENT"]
        SEO["🔍 SEO Optimization"]
        DESC["📝 Description
Generation"]
        CHAP["📑 Chapter
Markers"]
    end

    subgraph PlatformAdapters["🔌 PLATFORM ADAPTERS"]
        direction TB
        SPOT["🟢 Spotify
for Creators API"]
        YT["🔴 YouTube
Data API"]
        APPLE["🍎 Apple Podcasts
Connect"]
        AMAZON["📦 Amazon Music
RSS"]
        RSS["📡 Custom
RSS Feed"]
    end

    subgraph Analytics["📈 ANALYTICS"]
        LISTEN["👂 Listen Counts"]
        ENGAGE["💬 Engagement"]
        GROWTH["📊 Subscriber
Growth"]
    end

    AUDIO & META & ART --> SEO
    SEO --> DESC
    DESC --> CHAP
    CHAP --> SPOT & YT & APPLE & AMAZON & RSS
    SPOT & YT & APPLE & AMAZON --> LISTEN
    LISTEN --> ENGAGE
    ENGAGE --> GROWTH