> ## Documentation Index
> Fetch the complete documentation index at: https://docs.mirobody.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Mirobody Cloud

> Start with the hosted API, select an input for your health data, and choose an answer or agent workflow.

Mirobody Cloud provides the hosted `/v1` API for storing health data, standardizing readings, and answering questions over them. Your application collects the data and sends it to the service. For the product overview and the self-hosted path, see [Mirobody](/en).

## Getting started

<CardGroup cols={2}>
  <Card title="Run the first request" icon="rocket" href="/en/api-reference/quickstart">
    Create a key, write two sample readings, check their codes, and ask a question with `curl`.
  </Card>

  <Card title="Choose an API" icon="signs-post" href="/en/api-reference/choose-your-api">
    Select a grounded answer or an agent with tools, state, and streaming.
  </Card>
</CardGroup>

<h2 id="data-inputs">
  Data inputs
</h2>

Send each source to the endpoint that handles its shape. Use the same `user` value for every call that belongs to one Subject.

| Your data | Endpoint | Result |
| - | - | - |
| Device readings or other structured records | [`POST /v1/data`](/en/api-reference/data) | Stored readings with normalized names and units |
| Reports, photos, or spreadsheets | [`POST /v1/files`](/en/api-reference/files) | Stored source file and extracted readings |
| Narrative text or a preview of extraction | [`POST /v1/standardize`](/en/api-reference/extract) | Synchronous standardization result |
| A subjective journal entry | [`POST /v1/responses`](/en/api-reference/responses) with `store: true` | A stored conversation entry |

The [quickstart](/en/api-reference/quickstart) shows a complete structured-record request. The [API overview](/en/api-reference/overview) covers authentication, Subjects, errors, and retention.

## Processing stages

<Steps>
  <Step title="① Collect">
    Your application obtains device data, an uploaded report, or a user entry, then sends it through the matching endpoint above.
  </Step>

  <Step title="② Translate">
    The service matches measurement names to LOINC codes and parses units with UCUM. Inspect the resulting codes in the [standardization guide](/en/api-reference/standardization).
  </Step>

  <Step title="③ Agent">
    Ask about the stored record through the [Answers API](/en/api-reference/chat), or build an agent through the [Agent API](/en/api-reference/responses).
  </Step>
</Steps>

## Answer and agent APIs

| Workflow | Endpoint | Use it for |
| - | - | - |
| Grounded answer | [`POST /v1/chat/completions`](/en/api-reference/chat) | One question with supporting health records |
| Agent workflow | [`POST /v1/responses`](/en/api-reference/responses) | Tools, conversation state, and streaming |

Both use OpenAI-compatible request formats. For runnable Python and Node examples, see [SDK Examples](/en/api-reference/sdk-examples).

## Further reading

<CardGroup cols={2}>
  <Card title="API overview" icon="code" href="/en/api-reference/overview">
    Authentication, Subjects, retention, and the endpoint map.
  </Card>

  <Card title="Regions" icon="location-dot" href="/en/api-reference/regions/overview">
    Choose a console, key, and API host in the same region.
  </Card>
</CardGroup>