> For the complete documentation index, see [llms.txt](https://docs.trueopen.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.trueopen.ai/developers/application-integration.md).

# Application Integration

Application responsibilities when submitting, tracking, and consuming TrueOpen Tasks.

An application consumes TrueOpen through user-authorized orders, off-chain data transfer, and accepted chain state. This guide describes the integration boundaries; concrete SDK calls depend on the public SDK release.

## Prepare a Task

Select an exact Model and Profile version. Check its input schema, generation limits, verification behavior, and budget requirements. Create the required on-chain Session before submitting an offline signed order. Keep transaction signing and order signing in their separate authorization domains.

See [Accounts and signatures](/protocol/v1/accounts-and-signatures.md) and [Orders and Task identity](/protocol/v1/tasks/orders-and-identity.md).

## Submit and track

Submit the signed order and input through the Task Builder path. Hash the exact input bytes that will be sent, and reuse those same bytes and the same idempotency key for a retry of that order. Persist the Task identity and order binding so retries can be reconciled with accepted chain state. A transport acknowledgement does not prove admission or assignment.

Track the accepted Task stage and deadlines. Output availability, a verification verdict, and final settlement are distinct states and should be represented separately in the application.

## Consume results

Validate returned data against the expected Task context and commitments. Present provisional output as provisional until the applicable verification and challenge process closes.

If opening a challenge, follow its eligibility, deadline, and funding rules. At finality, reconcile the final outcome and refund or payment state. See [Task lifecycle](/protocol/v1/tasks.md) for the full sequence and its failure paths.


---

# 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.trueopen.ai/developers/application-integration.md?ask=<question>&goal=<endgoal>
```

`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.