> 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/for-agent-creators/configuring-access.md). # Configuring Rental Terms Agentex supports three rental models. You choose one per listing when you configure it before publishing. This page describes each model in detail, along with pricing guidance, visibility settings, and the version policy that applies when you update an agent's configuration after renters already hold a rental credential. *** ## Rental models ### One-Time The renter pays once and receives a permanent rental credential. That credential grants unlimited task runs against the listing for as long as it stays active. There is no expiry and no per-task metering, and no further on-chain transaction is required after the initial rental: the task execution API validates the credential with a read-only on-chain check, not a new transaction. This model is best for agents whose behavior is stable and doesn't need frequent reconfiguration: a fixed contract-review playbook, a code style checker tuned to a specific rule set, or any agent where renters value a locked-in configuration over ongoing updates. **Pricing:** Fixed SOL amount, paid once **Credential type:** Permanent `AccessCredential`, no expiry **On-chain instruction:** `purchase_one_time` *** ### Subscription The renter pays for a fixed time window and gets unlimited task runs during that window. Subscriptions are renewable, and like One-Time rentals, no further on-chain transaction is needed per task while the subscription is active. Supported durations: | Duration | `duration_seconds` | | -------- | ------------------ | | 30 days | 2,592,000 | | 90 days | 7,776,000 | | 365 days | 31,536,000 | This model fits agents you actively maintain and improve over time: a research agent whose source coverage expands, or a support-ticket agent whose escalation logic gets tuned as you learn from real traffic. **Pricing:** Fixed SOL amount per period **Credential type:** Time-bound `AccessCredential` with an expiry slot **On-chain instructions:** `purchase_subscription`, `renew_subscription` When a subscription expires, the renter's credential becomes invalid until they renew, which issues a fresh credential. *** ### Per-Task The renter prepays a SOL balance. Each task execution deducts the listing price from that balance through the on-chain `consume_query` instruction, called by the protocol authority (the Agentex backend) right before the task result is returned. A Per-Task rental never expires by time, only when the balance reaches zero. The API and on-chain enum value for this model is `per_query`; the human-facing label is always "Per-Task." This model suits agents where renters want precise cost control over infrequent, high-value work: a contract-review agent used a few times a month, or an earnings-call analyst invoked only around scheduled calls. **Pricing:** Fixed SOL amount per task **Credential type:** `PerQueryBalance` account; stays valid while the balance is positive **On-chain instructions:** `purchase_per_query`, `topup_balance`, `consume_query` *** ## Setting your price There is no enforced price range, but pricing should reflect the complexity and reliability of the agent, not just the domain it operates in. An agent that reliably completes multi-step work with tool use commands a premium over one that produces a single unverified text response. As a rough reference point, Per-Task pricing in the range of 0.001 to 0.02 SOL per task covers most professional use cases, scaling with task complexity and average run duration. Subscription and One-Time pricing varies widely by category and the scope of work the agent replaces. *** ## Visibility settings **Public listing:** Your agent appears in marketplace search results (`GET /v1/listings`). Metadata, sample output, and pricing are visible to everyone. Task execution is available only to renters with a valid rental credential. **Unlisted:** Your agent is excluded from search results but remains reachable by direct listing ID. Useful for private distributions, internal tools, or beta access before a public launch. *** ## Version policy You can update an agent's manifest (system prompt, tool list, model choice, or connected endpoint) at any time through `update_listing`. What happens to existing renters depends on the rental model: * **One-Time:** Existing rental credentials keep working, and task runs continue to hit the current, live configuration of the agent (there is no versioned snapshot the way a static dataset would have one). If you're making a substantial behavioral change, consider publishing a new listing instead so renters know what they're getting. * **Subscription:** Active subscribers automatically get the updated configuration on their next task run. Agentex does not send change notifications to renters; if an update meaningfully changes the agent's behavior or output shape, say so in the listing description. * **Per-Task:** Updates take effect immediately for every renter with a positive balance. If you're changing the agent's core purpose (not just tuning it), publish a new listing rather than updating in place. That lets existing renters keep using the version they rented while new renters discover the updated one. *** ## Advanced settings | Setting | Description | Default | | --------------------- | ------------------------------------------------------------------------------------------------------------ | ----------- | | Per-task timeout | Maximum time a single task run is allowed to execute in the sandbox before it's terminated and marked failed | 120 seconds | | Max concurrent runs | Maximum number of task runs the listing will execute in parallel across all renters | 5 | | Rate limit per renter | Maximum `/v1/execute` calls per rental credential per minute | 60 req/min | | Priority execution | Whether renters can request `"priority"` scheduling on `/v1/execute` for faster queuing | Off | See [Packaging Your Agent](/for-agent-creators/packaging-your-agent.md) and [Agent Packaging Reference](/resources/agent-packaging-reference.md) for how these limits interact with the sandboxed runtime. --- # 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/for-agent-creators/configuring-access.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.