Product

Agent flows

Three step-by-step views of elvaro in motion: how a question becomes an answer, how a signal lives and closes, and how personalization compounds over time.

1 · The life of a question

From the moment you press send to the streamed answer — every step, in order:

sequenceDiagram
    autonumber
    actor You
    participant C as Coordinator agent
    participant M as Market analyzer
    participant T as Trader agent
    participant D as Live data layer

    You->>C: "Should I take profit on SOL?"
    Note over C: classify intent,
load your trader profile
    C->>M: research market context
    M->>D: fetch live price + indicators
    D-->>M: SOL price, RSI, trend data
    M->>D: fetch sentiment / Fear & Greed
    D-->>M: sentiment reading
    C->>T: evaluate exit vs. your risk profile
    T-->>C: recommendation + concrete levels
    Note over C: synthesize evidence
into one answer
    C-->>You: streamed answer with reasoning

You see the progress live — fetching prices → running analysis → generating.

2 · The signal lifecycle

A signal isn't a message — it's a monitored object with a defined life:

stateDiagram-v2
    direction LR
    [*] --> Created: setup detected
    Created --> Active: entry level valid
(TP + SL attached)
    Active --> Monitoring: watched live,
24/7
    Monitoring --> TakeProfit: price hits TP ✓
    Monitoring --> StopLoss: price hits SL ✗
    Monitoring --> Invalidated: setup conditions
no longer hold
    TakeProfit --> Alerted: you're notified
(app + Telegram)
    StopLoss --> Alerted
    Invalidated --> Alerted
    Alerted --> [*]

Every exit path ends in an alert — you never find out hours later.

3 · The personalization loop

Your profile isn't set once — it compounds with every interaction:

flowchart LR
    A["🌍 Onboarding
conversation"] --> B["Trader profile
goals · risk · style"]
    B --> C["Personalized
answers & signals"]
    C --> D["Your activity
questions · bookmarks
positions you share"]
    D --> B
    E["⚙️ Settings
AI preferences"] --> B
    style B fill:#91e28c,stroke:#91e28c,color:#192824
    style C fill:#fffbcf,stroke:#f0ecc0,color:#192824

The loop: profile shapes answers, activity refines the profile.

Reading these diagramsThese flows are conceptual — they show the order and logic of what happens, not internal implementation. For the component view, see how elvaro works; for the terminology, see the AI behind elvaro.

Previous← Agent capabilities NextPersonalization & risk →