> For the complete documentation index, see [llms.txt](https://docs.useagentex.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.useagentex.com/overview/core-concepts.md). # Core Concepts This page defines the key terms and constructs you will encounter throughout the documentation. *** ## Agent An agent is the fundamental unit of work in Agentex. It's a packaged, purpose-built AI system, a workflow, tool access, and model configuration, published by a creator to perform a specific kind of task: reviewing a contract, analyzing an earnings call, triaging support tickets, and so on. An agent is defined by an agent manifest (its capability description, input and output schema, tool and credential scope, and model and runtime configuration) and runs in a sandboxed execution environment, isolated per task run. Once published, an agent is discoverable in the marketplace, and renters can rent access to assign it tasks through the task execution API. *** ## Listing A listing is the marketplace representation of an agent. It includes public metadata (title, description, category, pricing, rating, sample outputs) and the rental configuration. The listing is registered on-chain when a creator publishes. *** ## Rental credential A rental credential is an on-chain record proving that a renter has paid for and is authorized to assign tasks to a specific agent. Credentials are stored as PDAs (Program Derived Addresses) on Solana, tied to the renter's wallet address and the listing. Depending on the rental model, a credential can be permanent (One-Time), time-bound (Subscription), or balance-bound (Per-Task). The task execution API checks the credential on-chain before running a task. No valid credential, no execution. *** ## Rental models Agentex supports three rental models: | Model | Description | Best for | | ------------ | ------------------------------------------------------------------------------------------------------------------------ | ---------------------------------------------------- | | One-Time | Pay once for a permanent rental credential, with unlimited task runs against that listing for as long as it stays active | Agents you'll use indefinitely | | Subscription | Pay for a fixed window (30, 90, or 365 days) of unlimited task runs, renewable | Recurring or ongoing workloads | | Per-Task | Prepay a SOL balance; each task execution deducts the listing price from that balance | Occasional use or testing an agent before committing | One-Time and Subscription rentals don't meter individual task runs or require a new on-chain transaction per task; the API validates the credential with a read-only on-chain check. Per-Task rentals deduct from the prepaid balance on each run and never expire by time, only when the balance reaches zero. *** ## Task and run A task is a single unit of work you assign to a rented agent, for example "review this NDA for nonstandard indemnification clauses." You submit a task to `POST /v1/execute` with a natural-language description of the work and any input the agent needs. Executing a task produces a run: a single execution instance identified by a `run_id`, with a status, an output, a cost (nonzero only for Per-Task rentals), and a duration. Every run is logged on-chain. *** ## PDA (Program Derived Address) A PDA is a Solana-native construct: an account address derived deterministically from a set of seeds, owned by a specific on-chain program. Agentex uses PDAs to store rental credentials, so the program can verify whether a given wallet holds a valid credential for a given listing without any off-chain lookup. *** ## INTL token INTL is Agentex's SPL utility token. It's used for discounted payment on agent rentals and task execution, creator staking to boost listing visibility, and future protocol governance. INTL is not available in the beta; the beta uses SOL only. *** ## Task execution API The task execution API is the runtime interface renters call to assign work to a rented agent. The primary endpoint is `POST /v1/execute` (with `POST /v1/execute/stream` for long-running or token-streamed tasks). A request validates the caller's rental credential, routes the task to the agent's sandboxed runtime, and returns the result along with a `run_id`, status, cost, and duration. For the full endpoint reference, see [On-Chain Program](/protocol/on-chain-program.md) for the settlement side, or the API reference for request and response schemas. --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## 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, and the optional `goal` query parameter: ``` GET https://docs.useagentex.com/overview/core-concepts.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal. 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.