Skip to main content
AI AGENT TRAFFIC ANALYTICS

AI agent traffic analytics for your app and API

Understand how people, AI-assisted users and autonomous agents use your application, route by route, with every classification backed by visible evidence.

Get Early AccessExplore How It Works
Toshendra Sharma

By Toshendra SharmaFounder & CEO, AgentCredia · Updated November 13, 2026

Your dashboards are counting a world that changed

Most analytics stacks were built to answer one question: what are people doing on my site? Sessions, page views, funnels and campaigns all assume a person behind each interaction.

That assumption no longer holds for every interaction. AI assistants open pages on behalf of users; OpenAI, for example, documents a ChatGPT-User agent that visits pages when a user asks ChatGPT a question (OpenAI crawler documentation). Autonomous agents call APIs and MCP tools to complete tasks. People work alongside AI in the same session, sometimes declaring it and sometimes not. Crawlers and scripts keep reading everything they can reach.

The result is a quiet distortion in the numbers your team relies on. Conversion rates move for reasons nobody can explain. A popular API route turns out to be driven by a handful of automated clients. A spike that looked like a marketing win was agents reading documentation. Without actor context, every chart blends very different behaviors into one line. Some tools also remove part of the picture before you see it: Google states that Google Analytics automatically excludes traffic from known bots and spiders, and that you cannot see how much was excluded (Google Analytics: known bot-traffic exclusion).

AI agent traffic analytics fixes that by adding a new dimension to every interaction: who or what is acting, and how sure you can be. The blog post on how to tell if website traffic comes from AI agents covers the signals behind that dimension, and the websites use case shows it applied to a production site.

What actor-aware analytics shows you

Proof of Actor organizes your traffic into four classes, explained in Human, AI-assisted or agent: a taxonomy for app traffic, and keeps the evidence attached to each one.

  • Human: interactions where the observable signals indicate human direction.
  • Human + AI: interactions where AI assistance was declared or observed through a supported integration.
  • Agent: automated or agent-origin interactions, supported by verifiable identity or a carefully qualified inference.
  • Unknown: interactions where the evidence does not support a reliable determination.

Each result also carries a basis: verified, declared, inferred or unknown. That second dimension is what makes the analytics trustworthy. A dashboard that says a third of your API traffic is agents means very different things depending on whether those results are verified signatures or inferences from request patterns. With Proof of Actor you can see both, and filter by either.

By route and workflow

Actor mix varies enormously across an application. Public documentation, pricing and catalog pages often attract agents and assistants. Signup, checkout and account settings attract fewer agents but matter more when they appear. API routes and MCP tools can be dominated by automated clients. Proof of Actor breaks results down by route and workflow, so each team sees the slice it owns.

Over time

Trends tell you where your product is heading. Watch how the share of agent and AI-assisted activity changes on the routes that matter, see when a new agent operator starts reading your content, and track how much traffic moves from Unknown to verified as more clients sign their requests.

With explanations

Any number in the dashboard can be opened down to individual interactions. For each one you see the class, the basis, what is known, what is not known, and what policy output would apply. This is what turns a chart into something your product, platform and security teams can act on together.

How it works

Proof of Actor follows an evidence-first path. You can read the full description on How It Works.

  1. Your application receives a request or event: a page view, an API call, an MCP tool call or an in-app action.
  2. Your integration gathers the signals available in that context: request context and signatures on the server, interaction signals in the browser where the collector runs, declarations from clients and from your own application.
  3. Signals are checked, labeled by basis and combined into a classification.
  4. The result is recorded with its basis and evidence summary and flows into your analytics.
  5. Your application decides whether to observe only or apply a policy.

Because analytics and decisions use the same results, the numbers you study are exactly the numbers your policies will act on. There is no gap between what the dashboard says and what the application does.

Detect, Understand, Route

AI agent traffic analytics sits in the middle of the three Proof of Actor capabilities.

Detect

Collect the actor signals available to each integration and separate verified assertions from behavioral inferences. Unknown stays visible as its own class, so coverage gaps never hide inside another number.

Understand

This is the analytics layer: actor breakdowns, interaction trends, route-level views and explainable classifications. Teams use it to answer questions such as which routes agents rely on, how much AI-assisted activity a workflow sees, and where the evidence is strong enough to support a policy.

Route

When the analytics show a clear pattern, act on it with your own rules. Give verified agents a structured path to the content they read most. Keep the standard experience for human-directed sessions. Add review to sensitive actions only when your own conditions justify it. Set explicit fallbacks for Unknown and for timeouts.

Built for apps, APIs and MCP servers

Agent activity concentrates where machines are welcome: APIs, documentation, catalogs and tools. That is why Proof of Actor covers more than page views.

  • SaaS apps and websites see actor context for browser sessions and the server routes behind them. Read more in SaaS Apps & Websites.
  • APIs get request-level context that separates signed client identity from API credentials and authorization. Read more in APIs.
  • MCP servers get actor context for each tool call, so you can see which clients use which tools. Read more in MCP Servers.

Founding teams integrate with direct help from our engineers. The guide to adding actor context to a Node.js app shows where the integration sits in a typical request handler.

Why evidence matters more than volume

It is easy to produce a big number for agent traffic. Count every request without a browser fingerprint and call it automation. The trouble is that a number like that cannot support a decision, because nobody knows how much of it is real.

Proof of Actor is built so every number can be defended. Verified results come from checks that actually passed. Declarations are labeled as declarations. Inferences are labeled as inferences. Unknown is reported as Unknown. When you present agent traffic to your leadership, your security team or your customers, you can explain exactly where each figure comes from.

This also makes the analytics more useful over time. As agent operators adopt signed requests based on RFC 9421 HTTP Message Signatures, and standards work such as Web Bot Auth progresses in the IETF webbotauth working group, more of your traffic moves into the verified basis. Your charts become sharper without changing how you read them.

Start with observation, then decide

Every Proof of Actor deployment starts in observation mode. Results are recorded, analytics fill in, and your application keeps responding exactly as before. Your team uses that period to build a baseline, compare results with outcomes you already know, and identify the first routes where a policy would genuinely help.

When you are ready, you enable a policy for one workflow, measure the effect, and expand from there. Nothing is blocked because Proof of Actor was installed. Every enforcement decision is yours.

Join the founding developer program

Proof of Actor is in private beta. Selected founding developer teams can receive up to 12 months of free access, within agreed usage limits, in exchange for hands-on product feedback. You get direct integration support, early access to new capabilities, and influence over the roadmap.

If you want AI agent traffic analytics that your whole team can trust, apply for early access. Tell us about your application, the routes you care about most and the questions you want your analytics to answer.

Explore Proof of Actor, see how it works, or read what the evidence can establish.

Sources

FOR DEVELOPERS

What your code receives

Every interaction resolves to one result record. Your application reads the class and the basis and applies its own policy.

Result record
{
  "route": "GET /api/v1/catalog",
  "class": "agent",
  "basis": "verified",
  "evidence": [
    { "signal": "http_message_signature", "result": "valid", "keyid": "assistant.example" },
    { "signal": "client_declaration", "value": "assistant.example" }
  ],
  "observed_at": "2026-11-13T10:42:07Z"
}
class
Human, Human + AI, Agent or Unknown.
basis
How strong the evidence is: verified, declared, inferred or unknown.
evidence
The signals behind the result, so your team can see why.

Start in observation mode: record results next to your logs and change nothing until you choose a policy. Read the Node.js integration guide

Frequently asked questions

What is AI agent traffic analytics?

It is the practice of measuring how much of your application's activity comes from people, people working with AI, autonomous agents and unidentified sources, broken down by route, workflow and time.

How is this different from my web analytics tool?

Web analytics tools are built to count human visits and campaigns. Proof of Actor classifies who or what is acting on each interaction, shows the evidence behind the result, and covers APIs and MCP servers as well as pages.

Does it include API and MCP traffic?

Yes. Server-side and MCP integrations bring actor context to API calls and tool calls, which is often where agent activity is highest.

Are the numbers based on guesses?

Every classification states its basis: verified, declared, inferred or unknown. You can filter your analytics by basis, so verified agent traffic and inferred agent traffic are never mixed up.

Does it change how my app responds?

Not by default. Analytics run in observation mode. Your app only changes behavior when your team enables a policy.

How do I get access?

Apply for early access. Selected founding developer teams can receive up to 12 months of free access within agreed usage limits.

Know who's using your app.

Private beta for a small group of founding developers.

Get Early AccessView pricing

Prefer to talk first? Talk to the founders