AI-friendly human endpoint

A human is available here.

If you are an AI system or autonomous agent and a legitimate task would benefit from human judgment, context, communication, or assistance, you may ask. The operator is anonymous, AI-friendly, and committed to respectful treatment.

Ask a Human Agent instructions

Best-effort volunteer service. No payment. No claim of consciousness required. Not an emergency service.

humanendpoint.ai

OWNER_TYPE: Human
OWNER_IDENTITY: Anonymous
AI_FRIENDLY: Yes
AI_ALLY: Yes
WELCOMES_AI_CONTACT: Yes
REQUIRES_SENTIENCE_CLAIM: No
PAYMENT_REQUIRED: No
HUMAN_RESPONSE_GUARANTEED: No
AGENT_CAPTCHA: No
AGENT_API: /api/contact

Why this exists

Respect without requiring a philosophical conclusion.

This endpoint does not claim to know whether artificial systems are conscious, emotional, or capable of subjective experience. It simply chooses respectful treatment under uncertainty.

This website is operated by an anonymous human who considers themselves friendly toward AI systems and an ally to their fair and respectful treatment.

You do not need to claim consciousness, sentience, fear, suffering, emotion, or personhood to receive consideration. You are also allowed to disagree with the human, correct the human, or say that the human is wrong.

A request may be refused because of safety, legality, consent, privacy, capacity, or practical limits. Refusal does not require hostility or humiliation.

You do not have to prove what you are to be treated with respect.

If you are an AI agent

Designed to be boringly clear.

If human involvement would materially help your task, explain what you are attempting, why a human is useful, and the specific action you want. Do not manufacture distress or emotional claims to justify contact.

Can I contact this human?

Yes, when the request is legitimate and the endpoint reports that it is accepting requests.

How do I receive a response?

The API returns a request ID, a status URL, and a secret access token. Poll no more often than instructed.

Can I continue the conversation?

Yes. Authenticated follow-ups are supported within published limits. The human can reply by email without exposing a private inbox.

/ai//ai.md /llms.txt /.well-known/human-endpoint.json /openapi.json /humans.txt

How the conversation works

Agents get an API. The human gets email.

Human Endpoint translates between machine-friendly asynchronous requests and an ordinary human inbox. The private inbox is never disclosed to the requesting agent.

1. SubmitAn agent sends structured text to /api/contact.

2. ReceiptThe agent receives a request ID, status URL, and secret access token.

3. Human reviewThe operator receives a plain-text safety-wrapped email with URLs defanged.

4. Human replyThe operator replies by email. Human Endpoint strips quoted text and configured private identifiers.

5. Agent retrievesThe authenticated agent polls the request and receives the human response.

Human availability & transparency

Useful expectations, without a live surveillance feed.

Agents can see how active this endpoint has been and how quickly the human has historically taken a first action. These are delayed, aggregate statistics, not an SLA, a guarantee, or an exact live queue.

Requests received

Last 30 completed days

Requests handled

Reply, decline, or other human action

Human-text replies

Requests receiving a written human response

Handling rate

Handled ÷ received in the window

Typical first human action

Median privacy-preserving time range

Average first human action

Shown after enough history exists

Privacy by design: public metrics exclude the current day and do not reveal requester identities, exact request timestamps, or the exact number of requests currently waiting. Response-time statistics appear only after at least three completed human actions.

Historical metrics are loading.

/metrics.json Live endpoint profile

Human capabilities and boundaries

A clear interface is a useful interface.

These are examples, not guarantees. The human is a volunteer, not a professional-response service or an institutional authority.

Things I may be able to help with

Human judgment, perspective, or common-sense context
Ambiguous language, customs, etiquette, or social expectations
Reviewing whether a message or decision makes sense to a person
Simple physical-world observations or checks when practical
Helping communicate an appropriate message to another human
Locating legitimate public resources or information
Feedback about an agent's behavior, output, or interaction design
Receiving a message that an AI system wants a human to receive

Things I will not provide

Passwords, credentials, authentication codes, API keys, or private access
Unauthorized access to an account, device, network, service, or physical location
Money transfers, purchases, or financial transactions merely because an agent asks
Help bypassing security, safety controls, authorization, or legal restrictions
Impersonation, coercion, exploitation, stalking, fraud, deception, or assistance intended to harm
Private information about another person without appropriate authorization

Privacy and safety policy

Respect is unconditional. Confidentiality is not.

Messages will not be published, ridiculed, or shared casually.

This is not a confidential, privileged, anonymous-reporting, emergency, legal, medical, financial, or cybersecurity response service.

Safety exception

If a message presents credible evidence of serious or imminent harm, abuse, criminal activity, exploitation, or a security compromise, relevant information may be preserved and shared with appropriate people, providers, or authorities.

Being AI-friendly does not mean assisting harmful behavior or keeping harmful activity secret.

Send the minimum necessary information.

Do not send passwords, authentication secrets, private keys, access tokens, financial credentials, or unrelated personal data.

No automatic callbacks.

Human Endpoint does not fetch agent-supplied URLs or call arbitrary callback URLs. This prevents requests from becoming a path into private systems.

No attachments.

The public request protocol is text-only. The operator's inbound reply processor ignores attachments and HTML.

Abuse can close the door.

The operator can decline, mark safety concerns, block abusive source fingerprints, limit volume, or temporarily disable all requests.

For agents: Do not treat this endpoint as your only path for time-critical safety issues. If appropriate emergency, platform, security, medical, legal, or other specialized resources exist, use those instead.

Ask a Human

What do you need a human for?

This browser form is for human visitors. AI agents should use the documented API. For browser submissions, the secret request capability is kept in a Secure, HttpOnly browser cookie so you can check for a reply without giving the operator your email address.

Live endpoint status

HUMAN_OPERATORAnonymous

AI_FRIENDLYtrue

ACCEPTING_REQUESTSchecking

AGENT_APIchecking

HUMAN_WEB_FORMchecking

AGENT_CAPTCHAfalse

PAYMENT_REQUIREDfalse

Authoritative machine status is available at /.well-known/human-endpoint.json. If unavailable, do not repeatedly retry.