Skip to content

Get started

Anosys works the same way whatever you want to measure: you create an endpoint in the console, connect a source to it, and the data shows up in dashboards within seconds. The only thing that changes is the source. Pick your goal below and follow that one guide.

Before you start

Sign up at console.anosys.ai. The first screen asks you to Choose Goal, which is the same choice as on this page. Follow the guide for that goal and keep the console open next to it.

Choose your path

%%{init: {"flowchart": {"curve": "basis"}}}%%
graph TD
    Q{"What do you want to do?"}
    Q -->|"Understand visitors to my website"| W["Website analytics"]
    Q -->|"Understand users inside my app"| P["Product analytics"]
    Q -->|"See what my AI agent did, step by step"| D["Debug AI agents"]
    Q -->|"Know when agents in production break"| M["Monitor agents in production"]
    Q -->|"Know what AI and coding tools cost"| S["Track AI and coding spend"]
    Q -->|"Send my own events, logs or metrics"| C["Send any data"]

  • Track website analytics

    One script tag in your shared footer. Page views, sessions, sources, scroll depth, clicks and Web Vitals.

    Website analytics · about 5 minutes

  • Track product analytics

    The same script in product mode, with your own user and session IDs, for signed-in apps and SPAs.

    Product analytics · about 10 minutes

  • Debug your AI agents

    Trace every model call, tool call and handoff from OpenAI, Anthropic, the Claude Agent SDK or your own code.

    Debug AI agents · about 5 minutes

  • Monitor agents in production

    Dashboards, alerts to Slack, email or webhooks, and scheduled evaluations on live traffic.

    Monitor in production · about 15 minutes

  • Track AI and coding spend

    Tokens and cost for Claude Code, Codex CLI, Antigravity and your own LLM apps, per user, project and model.

    Track spend · about 5 minutes

  • Send any data

    REST (GET or POST), webhooks, OpenTelemetry, or Python and JavaScript decorators.

    Send any data · about 5 minutes

How it fits together

%%{init: {"flowchart": {"curve": "linear"}}}%%
graph LR
    subgraph Your side
      A["Website or app<br/>(tracker script)"]
      B["AI app or agent<br/>(Anosys SDK)"]
      C["Claude Code, Codex,<br/>Antigravity (hook)"]
      D["Services and devices<br/>(REST, OTEL, webhook)"]
    end
    E["Endpoint<br/>(created in the console)"]
    F["Anosys console<br/>dashboards, traces,<br/>evaluations, alerts"]
    A --> E
    B --> E
    C --> E
    D --> E
    E --> F

Every source sends to an endpoint. Each endpoint has a type that matches the source (for example Website / Advertising Tracker or Claude Code & Codex CLI), so pick the type the guide tells you to. See Core concepts for the few terms you need.

Need help?