<aside> 🧠

Client-facing explainer: This is the system I use to turn SEO research and ops into repeatable workflows. It combines an AI “operator” with scraping, research, and structured storage so work moves from question → evidence → decision → deliverable.

</aside>

What you get (in plain language)

• Faster answers with receipts. Competitor research, content gaps, and backlink prospecting based on actual source pages and data exports.
• Repeatable execution. Once a workflow works, it is packaged so it runs the same way every time.
• Clean outputs. Findings get written into a database or doc you can review, prioritize, and act on.

The architecture (system overview)

Integrations (what the system can plug into)

• Firecrawl: search, scrape, map, crawl, batch scrape, structured extract
• Google Drive: search/list, read files, create/copy, move/organize, share
• Notion: databases, pages, blocks, users (read + structured write-back)
• Parallel research: web search with domain filters, extraction from URL sets, async task runs
• Web search: real-time lookups when recency matters

<aside> 🔎

Want the full integration diagram + quick reference table? See: ‣

</aside>

The architecture (system overview)

flowchart LR
    subgraph AGENT["🤖 EXPERT SEO OPS AGENT"]
        direction TB
        AI["AI Core"]
        RUN["Run Tracker<br/>Speed Agent logic<br/>pl_run_output_retriever"]
    end

    subgraph WORKFLOWS["⚡ AUTOMATED FLOWS"]
        direction TB
        W1["🕷️ Jina AI Scraper"]
        W2["🔗 DataforSEO Backlinks"]
        W3["📥 Notion Backlink Writer"]

        W1 -->|"content"| W2
        W2 -->|"opportunities"| W3
    end

    subgraph SCRAPING["🔥 FIRECRAWL SUITE"]
        direction TB
        S1["scrape<br/>single URL"]
        S2["batch_scrape<br/>multiple URLs"]
        S3["crawl<br/>full site"]
        S4["map<br/>site structure"]
        S5["search<br/>web + scrape"]
        S6["extract<br/>LLM structured"]
    end

    subgraph RESEARCH["🔍 RESEARCH TOOLS"]
        direction TB
        R1["Parallel Search<br/>domain filters<br/>keyword queries"]
        R2["Parallel Extract<br/>content from URLs"]
        R3["Web Search<br/>real-time info"]
    end

    subgraph GDRIVE["📁 GOOGLE DRIVE"]
        direction TB
        G1["🔎 Search & List"]
        G2["📖 Read Files"]
        G3["✏️ Create & Copy"]
        G4["📂 Move & Organize"]
        G5["🔗 Share & Permissions"]
    end

    subgraph NOTION["📓 NOTION"]
        direction TB
        N1["🗄️ Databases<br/>query • filter • sort"]
        N2["📄 Pages<br/>search • retrieve"]
        N3["🧱 Blocks<br/>content extraction"]
    end

    subgraph CLIENT_DATA["📊 CLIENT WORKSPACE"]
        direction TB
        D1[("SEO Docs<br/>Drive Folder")]
        D2[("SEO Spreadsheet<br/>Rankings Data")]
        D3[("Notion Hub<br/>Tasks • Calendar<br/>Keywords • Backlinks")]
    end

    AI <--> WORKFLOWS
    AI <--> SCRAPING
    AI <--> RESEARCH
    AI <--> GDRIVE
    AI <--> NOTION
    RUN -.->|"status + routing"| WORKFLOWS

    GDRIVE <-.-> D1
    GDRIVE <-.-> D2
    NOTION <-.-> D3

    style AGENT fill:#0d1117,stroke:#58a6ff,stroke-width:3px,color:#fff
    style WORKFLOWS fill:#161b22,stroke:#f78166,stroke-width:2px,color:#fff
    style SCRAPING fill:#161b22,stroke:#ff7b72,stroke-width:2px,color:#fff
    style RESEARCH fill:#161b22,stroke:#d29922,stroke-width:2px,color:#fff
    style GDRIVE fill:#161b22,stroke:#3fb950,stroke-width:2px,color:#fff
    style NOTION fill:#161b22,stroke:#a371f7,stroke-width:2px,color:#fff
    style CLIENT_DATA fill:#21262d,stroke:#8b949e,stroke-width:2px,color:#fff

Speed Agent logic (how this stays fast and reliable)

This system uses Speed Agent logic to keep work moving without wasting cycles:

• Fast-path routing: choose the shortest workflow that can answer the question.