Cloudflare sits in front of a large share of the web. For many teams it is the CDN, the DNS provider, the firewall and the first line of defence against abusive automation, all in one place. Its bot products are designed for that position: they see traffic early, decide whether it should be challenged or stopped, and give site owners control over how crawlers, including AI crawlers, reach their content.
Your application has a different question to answer. Once a request is through the edge, your code needs to know who or what is acting so it can choose the right experience. Is this a person, a person working with an AI assistant, an agent acting for a customer, or something the evidence cannot explain? And how sure should your code be?
That is the job Proof of Actor is built for. It is a focused, application-level layer that returns an actor class with an explicit evidence basis, delivered to the code that owns the decision. This page compares the two fairly, using Cloudflare's own documentation as of November 2026, so you can see where each belongs.
The short version
| Cloudflare bot products | Proof of Actor | |
|---|---|---|
| Built for | Protecting infrastructure and content at the edge | Giving your application actor context for product and policy decisions |
| Where it runs | Cloudflare's edge network, in front of your origin | Your server, edge layer, approved browser pages, HTTP API or MCP server |
| Main output | Challenges and rule actions; for Enterprise Bot Management, a bot score from 1 to 99 and bot fields | Human, Human + AI, Agent or Unknown, plus a basis: verified, declared, inferred or unknown |
| Where the output lands | Cloudflare rules, analytics and Workers | Your application code and actor analytics by route and workflow |
| AI-assisted people | Not a separate category in the bot score | Their own class, Human + AI, when declared or observed via a supported integration |
| Unclear traffic | Scored and handled by edge rules | Unknown as a first-class result with a fallback you set per route |
| Signed agents | Web Bot Auth verification and signed agent signals | RFC 9421 signature checks reported as a verified basis in your code |
| MCP servers and API callers | Protected as HTTP traffic at the edge | Caller context inside the server that handles the tool call or API request |
| CDN dependency | Applies to traffic that passes through Cloudflare's network | Independent of which CDN, if any, you use |
| Default on install | Depends on the product and settings you enable | Observation mode; nothing changes until you add a policy |
The rest of this page explains each row.
Looking for a Cloudflare Bot Management alternative for AI agent traffic?
Most teams searching for an alternative are not unhappy with the edge. They have a new question the edge was not designed to answer: an AI agent, an AI-assisted person and an ordinary visitor are all reaching the same signup form, API route or MCP tool, and the product needs to treat them differently. Here is a plain split.
Where Proof of Actor replaces the decision. If you currently use bot scores or bot fields to drive product behaviour, for example labelling AI-drafted submissions, routing agents to a structured endpoint, or choosing a fallback for unclear callers on a specific workflow, Proof of Actor is the better fit for that job. It returns an actor class, Human, Human + AI, Agent or Unknown, with an evidence basis, directly to your application code. The routing and policy decision moves into the code that already knows what each route does, and it works the same way whichever CDN you use.
Where Proof of Actor adds to Cloudflare. Proof of Actor is not a network security product. According to Cloudflare's documentation as of November 2026, its DDoS protection automatically detects and mitigates attacks and is available on all plans, and its Web Application Firewall checks incoming web and API requests against managed and custom rules. Keep that layer. Volumetric attacks, exploit attempts and abusive automation should be stopped before they reach your origin, and Proof of Actor does not try to do that.
So the direct answer to "is this an alternative?" is: yes for actor classification and per-route routing in your app, and no for edge DDoS and WAF protection. Most production teams end up with both layers, each doing the job it was built for.
What Cloudflare's bot products are built for
According to Cloudflare's documentation as of November 2026, it offers three bot-specific products:
- Bot Fight Mode, available on the Free plan, which challenges detected bot traffic across a domain with a single toggle.
- Super Bot Fight Mode, for Pro, Business and Enterprise plans without the Bot Management add-on, which adds configurable actions per bot category.
- Bot Management for Enterprise, which adds per-request bot scores, custom rules, per-endpoint handling and detailed analytics.
The bot score is, in Cloudflare's words, a score from 1 to 99 that indicates how likely a request came from a bot, where 1 means highly confident the request was automated and 99 means highly confident it came from a human. Cloudflare's documentation states that granular scores are available to Enterprise customers who have purchased Bot Management, while other plans see grouped results such as automated, likely automated, likely human and verified bot. Scores can be used in WAF custom rules to block, challenge or allow requests, and they are also available in Workers.
Cloudflare also maintains verified bots, which its documentation describes as bots or agents that are transparent about who they are and what they do, such as search engine crawlers, monitoring services and user-driven agents. Verification relies on methods such as a Web Bot Auth signature or IP validation, plus non-abusive behaviour.
For AI traffic specifically, AI Crawl Control gives site owners visibility into which AI services access their content, lets them allow or block individual crawlers, and tracks how well crawlers follow robots.txt. Cloudflare's documentation lists it as available on all plans.
These are strong tools for the job they are designed for: deciding at the edge which automated traffic reaches your origin and your content.
Where each runs, and where its output lands
The most practical difference is location.
Cloudflare's bot products run on Cloudflare's network, in front of your origin. Their output lands where Cloudflare's controls live: WAF custom rules, bot analytics, and, according to Cloudflare's bot management variables reference, the request.cf object in Cloudflare Workers. That is the right place to decide whether a request should be stopped or challenged.
Proof of Actor runs where your application logic lives: Node.js middleware, Next.js, the edge layer you control, a browser collector on pages you approve, a generic HTTP API, or your MCP server. Its output lands in your application code as a structured result, and in actor analytics broken down by route and workflow. See the integration paths for details.
That location matters for product decisions. Your code knows which workflow a route belongs to, whether your own assistant performed an action, whether a client declared itself, and what the business impact of that route is. A per-route decision such as "serve a verified agent the structured endpoint, keep people on the normal flow, label AI-assisted submissions" belongs next to that knowledge. Proof of Actor puts the actor context exactly there, and because it runs in your stack rather than in one provider's network, it works the same behind Cloudflare, another CDN, or no CDN at all.
Scores versus a class with a basis
A score is an efficient input for a protection rule: choose a threshold and act above or below it. It is designed to answer how likely a request is to be automated.
Product decisions need a richer answer. Proof of Actor returns one of four classes:
- Human: observable signals indicate human-directed interaction.
- Human + AI: a person directed the interaction with AI assistance that was declared or observed via a supported integration.
- Agent: automated or agent-origin interaction.
- Unknown: the evidence does not support a reliable determination.
Every result also carries its evidence basis: verified, for example an RFC 9421 HTTP Message Signature checked against a published key; declared, when a client or your application stated it; inferred, clearly labeled as inference; or unknown. Your team can read each result and see the evidence behind it, which is what you need when the decision affects a customer's experience rather than an attack surface.
AI agents, signed agents and AI-assisted people
Cloudflare has invested in agent verification. According to its documentation as of November 2026, Cloudflare supports Web Bot Auth, an authentication method that uses cryptographic signatures in HTTP messages, based on RFC 9421, to verify bots and agents. Its docs describe a signed agent signal, exposed as the cf.bot_management.signed_agent field for Enterprise customers with Bot Management, and from July 1, 2026 a Direct versus Intermediary access field that distinguishes single-operator bots from agentic services many end users can operate. Cloudflare also offers AI crawler and agent access policies across plans.
Proof of Actor treats signed requests as one part of a wider picture. When a request carries a valid HTTP Message Signature checked against a published key, the result is Agent with a verified basis, delivered straight to your route handler. The guide on verifying signed agent requests shows the flow. A valid signature proves which key signed the request, not what that agent is authorized to do, so authorization stays in your application where it belongs.
The bigger difference is the people in the middle. Some of your users work with an AI assistant beside them: drafting support tickets, filling forms, moving through onboarding. Edge bot scoring is designed to separate automation from people, and it has no reason to carve out this group. Your product often does. Proof of Actor reports Human + AI as its own class, only when the assistance is declared or observed through a supported integration, so you can label AI-drafted submissions, measure how AI-assisted users move through a workflow, or give them a smoother path. Read more in the AI-assisted interaction glossary entry.
Unknown is a result, not a threshold
Every detection system meets traffic it cannot explain. At the edge, unclear traffic is handled by whatever rule the score falls under. That keeps protection consistent across a whole zone.
Inside an application, one setting rarely fits every route. An unclear request to a public documentation page and an unclear request to change account credentials deserve different handling. Proof of Actor returns Unknown as a first-class result, tracks your Unknown rate over time, and lets you choose the fallback per route: allow it on public content, keep existing checks on account changes, send it to review on high-impact actions. The guide on handling Unknown traffic covers the patterns, and the Unknown classification entry explains why an honest Unknown beats a forced guess.
MCP servers and APIs
Cloudflare protects MCP servers and APIs the same way it protects any HTTP traffic that passes through its network. That is valuable, and it happens before the request reaches your service.
What an MCP server or API usually needs next is caller context inside the handler: which agent is calling this tool, is that identity verified or only declared, and should this tool call be allowed, adapted, rate-limited or sent for review. Proof of Actor ships an MCP server integration and a generic HTTP API so that context arrives where the tool or endpoint logic runs. See MCP server caller identification and the APIs use case for examples.
How to use both together
The strongest setup gives each tool its own layer:
- At the edge, Cloudflare absorbs attacks, challenges abusive automation, verifies known bots and applies your AI crawler policies.
- In your application, Proof of Actor classifies the traffic that reaches you as Human, Human + AI, Agent or Unknown, with a basis.
- In your code, per-route policies use that context to allow, adapt, challenge, send for review or rate-limit, with an explicit fallback for Unknown.
- In your analytics, product and security teams see which routes and workflows each actor class uses.
Proof of Actor starts in observation mode, so adding it behind Cloudflare changes nothing on day one. You see your real actor mix, preview a policy against live traffic, and enforce one route at a time. The guide on rolling out actor-aware policies safely walks through it. Proof of Actor performs no identity checks, KYC or biometrics; it reports who or what is acting and leaves the decision to you.
When to choose which
Choose Cloudflare's bot products when your main need is edge protection: stopping abusive automation before it reaches your origin, challenging suspicious traffic across a zone, and controlling how AI crawlers access your content.
Choose Proof of Actor when your main need is application-level actor context: knowing whether people, AI-assisted people or agents use each route, seeing the evidence behind every result, handling Unknown deliberately, giving MCP server and API handlers caller context, and writing per-route product policies that work regardless of your CDN.
If you run a production website, API or MCP server, you likely need both answers, which is why the two sit comfortably in the same stack. For the broader category comparison, see Proof of Actor vs bot management.
Get started
Proof of Actor is in private beta. Through the founding developer program, selected teams can receive up to 12 months free within agreed usage limits, with no automatic billing. Apply for early access and tell us which application, API or MCP server you want to understand first.
Product details about Cloudflare on this page are based on Cloudflare's public documentation, including the bot score reference and the Web Bot Auth reference, the DDoS protection overview and the WAF overview, as of November 2026.
Cloudflare is a trademark of Cloudflare, Inc.