
Proof of Actor by AgentCredia
Know who's using your app: people, people with AI, or AI agents
About Proof of Actor by AgentCredia
Proof of Actor is AgentCredia's tool for telling apart who is acting in a web app, API or MCP server: a person, a person using AI, an autonomous AI agent, or unknown. Each interaction gets a class plus the strength of the evidence behind it, and your team decides what the app does with that. AgentCredia is run by Recordskeeper Inc, with an office in India. The product is in private beta: every plan, including Free, starts with an early access application, and integrations are set up hands-on with each team. It is aimed at developers running SaaS apps, websites, public APIs and MCP servers.
You add a supported integration: Node.js or Next.js middleware, a browser script, a Cloudflare edge integration, a generic HTTP API or an MCP server integration. It collects signals such as signed agent requests, client declarations and browser interaction signals, labels each as verified, declared, inferred or unknown, and returns a result record with the class and the evidence. Everything starts in observation mode, which changes nothing in your app. When you trust the results, you set per-route policies to allow, adapt, challenge or limit specific actors.
- •Four actor classes - Every interaction is labelled Human, Human + AI, Agent or Unknown, and Unknown stays a separate result instead of being folded into a guess.
- •Evidence basis on each result - Each result says whether it rests on a verified signature, a client declaration, a behavioural inference or nothing, with the signals listed.
- •Observation mode first - Installing it blocks nothing; results are recorded next to your logs so you can review false positives before enforcing any rule.
- •Per-route policies - Rules by route, actor class and evidence basis, such as sending verified agents to a structured API or asking for a stronger credential.
- •Actor mix dashboard - Shows the share of people, AI-assisted users, agents and unknown callers per route, and which signed agents call your app.
SaaS and website teams use it to see how much of their traffic is AI agents and where agents scrape pages built for people. API and MCP server operators use it to identify callers before high-impact actions. Product and security teams use the policy preview to plan agent access rules without blocking automation their customers rely on.
Pricing
$0 - $99/mo
- Free$0/mo
- Pro$99/mo
- EnterpriseContact sales
























































