# Bright Data search_engine: taming AI overviews and token-heavy results

## TL;DR

If search results are bloated or contaminated for agent use: 1. Upgrade to MCP v2.8.0 or later, which added structured output and control over the extra result decorations. For high-precision agent research, disable the AI overview and image scoring; they cost a lot of tokens and pollute exact searches, while casual users may prefer them.

Use this when you run into the situation in the title.

**When not to use this skill:** unrelated tasks. It covers only the procedure above.

## Compatibility

The steps above apply to the commands named in them. This skill does not pin a version, so if a flag looks different on your machine, check your installed version's docs first.

## Details

2. 3. Expect the remote (hosted) version to get output-format changes before the STDIO version; if a feature seems missing locally, check whether it is live remotely first.

Context: GitHub issue (brightdata/brightdata-mcp#66, closed): the reporter wanted control over what the search_engine tool returns because AI overviews and image scoring were contaminating high-precision agent searches and burning tokens for no reason. The maintainer pointed at PR #57 as the structural fix and released v2.8.0 with the fix; structured output went live for the remote version first, then STDIO. The reporter's reaction: enthusiastic confirmation that this saves a lot of unnecessary tokens.