Humalike
Humalike is a behavioral API platform for agents that need to participate in human conversations with better timing, persona, social memory, social learning, social observability, and theory-of-mind style response checks.
Many agents can answer questions but still feel wrong inside group chats, communities, voice flows, tutoring, NPCs, or coworker-style workflows. Humalike matters because it turns social behavior into explicit APIs, docs, billing, research artifacts, and a Hermes plugin instead of leaving it as prompt-only product polish.
Humalike official docs describe a suite of HTTP APIs for agents that talk, time, and read people like humans, with bearer-token authentication, JSON requests and responses, credits, and named API groups. The official site lists seven behavioral components, use cases, research links, trust notes, and a $20 free-start credit claim. Product Hunt launch material and the maker post add demand language and seven-API positioning, while arXiv and the Hermes plugin provide supporting research and integration context.
- Add turn-taking and conversation pacing to group-chat agents.
- Give agents person-centric memory inside a shared conversation scope.
- Analyze whether an agent reply will land poorly before sending it.
- Improve NPCs, tutors, AI coworkers, companions, communities, and humanoid workflows where social timing matters.
Humalike docs describe a shared API platform around social behavior. The core API families include turn-taking, persona, social learning, social observability, social memory, and theory of mind. The reader job is not only "what model is this?" but "which social behavior should my agent outsource to an API?"
- Turn-taking: decide when an agent should speak and pace replies.
- Social memory: ingest, recall, and ask over person-centric group conversation memory.
- Social observability and theory of mind: inspect how a message lands and refine behavior before sending.
Humalike is most relevant when an agent enters a shared human environment: group chats, online communities, classrooms, games, hardware companions, customer support, or humanoid workflows. In those settings, answer quality is not enough; timing, memory, social norms, and escalation behavior become part of product quality.
The official site says SOC 2 Type II and ISO 27001 are in progress and describes encryption at rest and in transit. The docs say billable calls are metered in credits, priced before running, and rejected with `402 Payment Required` when the account balance cannot cover them. Product Hunt comments and research links are useful context, but official docs should control implementation facts.
Humalike is a behavioral service a harness can call during an agent loop.
Companion-style products face similar timing, memory, and status-expression questions.
Conversation memory and behavioral APIs need clear data, permission, and billing boundaries.
Source confidence
Humalike
Humalike Docs
Humalike Docs
Humalike Docs
arXiv
GitHub / Humalike
Product Hunt
Humalike FAQ
Page-level questions for Humalike.
What problem does Humalike solve for AI agents?+
Humalike targets the social behavior gap: agents that are capable but awkward in conversation. Its APIs help with when to speak, how to adapt tone, what to remember about people, how to read group norms, and how a message might land before it is sent.
Is Humalike only for voice agents?+
No. Product Hunt labels it partly as voice-agent infrastructure, but the official site and docs describe broader use cases: group chats, communities, games, coworkers, tutors, companions, care workflows, and humanoids. The common thread is social behavior, not one interface.
How does Humalike charge for API calls?+
Humalike uses account credits for billable calls. Its billing docs say requests are priced before work begins, capped at that price, failed calls cost zero, and underfunded requests return `402 Payment Required` before running.
What should teams verify before using Humalike?+
Verify which APIs are available for your workflow, what data you will send, whether the credit model fits expected traffic, and what trust requirements apply. The official site says SOC 2 Type II and ISO 27001 are in progress, so teams with strict compliance needs should confirm current status directly before production use.