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

# Fast mode

> Search mode optimized for speed, returning results in ~1 second.

Fast mode returns results in about 1 second. The default `auto` mode already resolves simple queries fast — pin `searchMode=fast` when you need that latency guaranteed on every call, such as real-time applications and high-volume workloads. Equivalent to [`effort=low`](/search/search-modes#the-effort-parameter), if you use the generic effort dial instead of naming a mode.

<CodeGroup>
  ```bash curl theme={null}
  curl "https://api.andiai.com/api/v1/search?q=latest+AI+news&searchMode=fast" \
    -H "x-api-key: YOUR_API_KEY"
  ```

  ```python Python theme={null}
  import requests

  response = requests.get(
      "https://api.andiai.com/api/v1/search",
      params={"q": "latest AI news", "searchMode": "fast"},
      headers={"x-api-key": "YOUR_API_KEY"}
  )

  data = response.json()
  for result in data["results"]:
      print(f"{result['title']} — {result['source']}")
  ```

  ```javascript JavaScript theme={null}
  const response = await fetch(
    "https://api.andiai.com/api/v1/search?q=latest+AI+news&searchMode=fast",
    { headers: { "x-api-key": "YOUR_API_KEY" } }
  );

  const data = await response.json();
  data.results.forEach(r => console.log(`${r.title} — ${r.source}`));
  ```
</CodeGroup>

## When to pin fast mode

* Real-time search in user-facing applications with a hard latency budget
* High-volume automated queries
* Applications where latency matters more than exhaustive coverage

For most applications, omit `searchMode` and let `auto` decide — it resolves simple queries fast on its own. See [search modes](/search/search-modes) for the full set.

## Example response

```json theme={null}
{
  "results_type": "News",
  "answer": "",
  "type": "News",
  "title": "latest AI news",
  "results": [
    {
      "title": "Latest AI News and Developments",
      "link": "https://example.com/ai-news",
      "desc": "A roundup of the latest developments in artificial intelligence...",
      "source": "example.com",
      "date": "2026-07-14T10:00:00.000Z"
    }
  ],
  "metrics": {
    "query": "latest AI news",
    "intent": "LatestNewsIntent",
    "timestamp": "2026-07-14T21:30:00.000Z",
    "duration": 890,
    "queries_executed": 1,
    "api_requests_count": 1,
    "results_returned": 10,
    "total_results_found": 50,
    "cost_dollars": 0.0031
  }
}
```

### Parsing the response

```python theme={null}
data = response.json()

# Check what type of results came back
print(data["results_type"])  # e.g., "News", "Search", "Weather"

# Access results
for result in data["results"]:
    print(result["title"], result["link"])

# Check performance and cost
print(f"Returned {data['metrics']['results_returned']} results in {data['metrics']['duration']}ms")
print(f"Cost: ${data['metrics']['cost_dollars']}")
```

<Accordion title="Adding extracts and metadata">
  Get richer data from each result:

  ```bash theme={null}
  # Text extracts from result pages
  curl "https://api.andiai.com/api/v1/search?q=latest+AI+news&extracts=true" \
    -H "x-api-key: YOUR_API_KEY"

  # Full metadata including content type and reader data
  curl "https://api.andiai.com/api/v1/search?q=latest+AI+news&metadata=full" \
    -H "x-api-key: YOUR_API_KEY"

  # Markdown format for passing to LLMs
  curl "https://api.andiai.com/api/v1/search?q=latest+AI+news&format=context" \
    -H "x-api-key: YOUR_API_KEY"
  ```

  <Warning>
    `metadata=full` adds latency because it fetches additional data from each result page. Use `metadata=basic` (the default) unless you need `contentType` or `reader` data.
  </Warning>
</Accordion>

## Fast mode vs. deep mode

|                  | Fast mode      | Deep mode                |
| ---------------- | -------------- | ------------------------ |
| Response time    | \~1 second     | \~2–3 seconds            |
| Spell correction | Detection only | Corrects and re-searches |
| Topic coverage   | Single angle   | Multiple angles          |
| Result quality   | Good           | Thorough                 |

<Tip>
  Start with `auto` — it picks between these per query. Pin fast mode when latency is critical; pin deep mode when thoroughness matters more than speed. See [search modes](/search/search-modes) for the full set of modes.
</Tip>

## Next steps

<CardGroup cols={2}>
  <Card title="Deep mode" icon="microscope" href="/search/deep-search">
    Multi-angle search with spell correction.
  </Card>

  <Card title="Search modes" icon="gauge-high" href="/search/search-modes">
    Automatic effort by default, manual control when you want it.
  </Card>

  <Card title="Basic search example" icon="code" href="/examples/basic-search">
    Complete integration with error handling.
  </Card>

  <Card title="Response format" icon="brackets-curly" href="/features/response-format">
    Response structure and metrics.
  </Card>
</CardGroup>