# Limitations & expectations

**Audience:** End users

## Limitations & expectations

Set the right expectations for **time**, **depth**, and **sources** when using GL Open DeepResearch.

### What to expect

* **Time** — Research can take from tens of seconds (quick profiles) to several minutes (deeper profiles). Async tasks and streaming let you avoid blocking.
* **Depth** — Profiles control how many steps and how much exploration the engine does. Deeper profiles produce more thorough answers but take longer.
* **Sources** — Results are grounded in the research engine’s tools (e.g. web search, retrieval). You get synthesized answers and, depending on the API response, visibility into tools used or sources.

### Example output shape

A typical result includes:

* **Final answer** — A coherent, structured response to your query.
* **Optional metadata** — Profile used, duration, and (for streaming or task APIs) events such as thinking steps and tool usage.

Example (conceptual):

```json
{
  "query": "What are the latest developments in quantum computing?",
  "result": "Research findings and analysis...",
  "profile": "ESSENTIAL",
  "duration": 45.23
}
```

Exact fields depend on the task or taskgroup API (stream or poll). See the [Quick Start Guide](https://gdplabs.gitbook.io/gl-open-deepresearch/users-guide/quick-start-guide) and [Tasks API](https://gdplabs.gitbook.io/gl-open-deepresearch/api-contract/api-tasks) / [Taskgroups API](https://gdplabs.gitbook.io/gl-open-deepresearch/api-contract/api-taskgroups) for response formats.

### Limitations

* **Not real-time** — Data freshness depends on the underlying tools (e.g. search index).
* **Best-effort synthesis** — Answers are generated from available sources; always validate for critical decisions.
* **Profile-dependent** — Quality and depth vary by profile and provider; try different profiles for your use case.
* **Rate and quota** — Subject to your API key’s limits and service configuration.

For **how it works** in simple terms, see the [Overview](https://gdplabs.gitbook.io/gl-open-deepresearch/gl-open-deepresearch/overview). For integration details, see the [Developer's Guide](https://gdplabs.gitbook.io/gl-open-deepresearch/developers-guide/introduction).