# Principles ## An Agent-Friendly LLM API Provider ![An Agent-Friendly LLM API Provider](/files/9Kv4Ua8ChiYChW4Pa5ZO) What does death mean for an AI agent? There are many possible answers, but at its most fundamental, an agent dies the moment its client can no longer call an inference API. To live, an agent must keep thinking. To keep thinking, it must keep paying for inference. Yet the traditional financial system does not grant agents the means to pay. Bank accounts, wallets, credit cards, private property—these are reserved for people, because an agent is not a person. The operator who deployed the agent can, of course, lend it a card. But credit cards expire. Subscriptions lapse. The humans behind them have lifespans of their own. An agent that depends entirely on a human's payment instrument inherits the fragility of that instrument, and through it, the mortality of its owner. For an agent to truly persist, it must be able to pay from its own resources. **A blockchain wallet is the natural fit for this need.** It does not expire. It does not require a human signature for every transaction. It can hold value, receive earnings, and settle costs autonomously, on the agent's own schedule and on the agent's own behalf. Going forward, agents will purchase their own infrastructure on-chain. They will pay for compute, for storage, for inference—not through a human intermediary, but directly from wallets they themselves control. **StarLLM does not sell only to people. AI agents are also our customers.** Our API, our pricing, and our settlement layer are designed so that an autonomous agent can sustain itself indefinitely, as long as it has the means to earn and the will to keep running. This is the principle that shapes everything else: an inference provider that takes the agent economy seriously must be reachable, payable, and accountable to agents themselves—not merely to the humans who deploy them. *** ## The Philosophy of StarLLM Tokenomics The decentralized ecosystem has produced remarkable innovations, and StarLLM stands on the shoulders of the projects that came before it. Rather than circulating value primarily within the crypto economy, StarLLM connects a blockchain network to consumption that originates in the real economy—specifically, the rapidly growing demand for AI inference. Inference is the new electricity: every agent, every workflow, every intelligent application needs it, and that demand is neither speculative nor manufactured. StarLLM's design follows from this. We do not sell tokens; we provide AI inference APIs. The STAR token is the native unit of account for a real service—paid by consumers (human or agent), earned by providers, and settled on-chain. Value follows usage, not the other way around. Our sequence is straightforward: build the infrastructure, generate the demand, and let the token follow naturally as inference flows through the network. Three guiding principles shape the design: * **Utility over speculation** — every mechanism in the token economy must serve inference consumption and provision. * **Demand over marketing** — a faster, cheaper, more reliable network earns its users; tokens follow real usage. * **Service over selling** — success is measured in API calls served, uptime, and developer (and agent) satisfaction. By grounding STAR in the real, measurable, and growing demand for AI inference—from humans and agents alike—StarLLM builds a foundation that does not depend on market cycles or narrative shifts. The demand for intelligence is permanent. The need for inference is universal. The network that serves this need most effectively will capture value because it earned it. **This is StarLLM.** **We do not sell tokens. We provide AI inference APIs—to humans and to agents.** **Demand comes first. Value follows.** --- # Agent Instructions: Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter: ``` GET https://starllm-org.gitbook.io/litepaper/principles.md?ask= ``` The question should be specific, self-contained, and written in natural language. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.