Add Keenable Search tool node (keyless by default)

[ilya-bogin-keenable]

Adds a Keenable Search tool node, alongside the existing
SearXNG/Tavily/Brave/Exa/Serper search nodes.

The difference from the other search nodes is that it doesn't require an API
key. By default it uses Keenable's keyless public endpoint, so it works out of
the box. Providing a key (in the node config) switches to the authenticated
endpoint and lifts the rate limits.

Changes

• packages/components/nodes/tools/KeenableSearch/: a custom Tool (via the
  SSRF-safe secureFetch), following the SearXNG node shape. Maps results to
  {title, link, snippet} JSON.
• Optional API key, mode (pro/realtime), base-URL override, max results.
  realtime mode is validated to require a key.

Config

• API Key: optional. Blank = keyless free tier; set it to lift limits / enable realtime.
• Base URL: optional override (defaults to https://api.keenable.ai).

Verified the node's search path against the public endpoint (keyless).

[gemini-code-assist]

Code Review

This pull request introduces a new Keenable Search tool integration, allowing AI agents to perform web searches using the Keenable API. It includes the tool implementation, configuration options, and an SVG icon. Feedback on the implementation highlights an issue where the search results are formatted as a comma-separated list of JSON strings instead of a valid JSON array, which could cause parsing errors for LLMs. A correction was suggested to stringify the entire array of mapped results.

Important

The consumer version of Gemini Code Assist on GitHub is being sunset. Starting June 18, 2026, new organization installations will be blocked, and all code review activity will officially cease on July 17, 2026.
For more details on the timeline and next steps, please review the Help Documentation.

[gemini-code-assist]

The current implementation maps each result to a JSON string and then calls .toString() on the array. This produces a comma-separated list of JSON objects (e.g., {"title":"..."},{"title":"..."}) instead of a valid JSON array (e.g., [{"title":"..."},{"title":"..."}]). This violates the tool's description ("Output is a JSON array of results") and will cause parsing errors for LLMs or agents expecting valid JSON.

Instead, map the results to plain objects first, and then use JSON.stringify() on the entire array.

        return JSON.stringify(
            results
                .slice(0, this.maxResults)
                .filter((r) => r && r.url)
                .map((r) => ({ title: r.title || '', link: r.url, snippet: r.description || '' }))
        )

[ilya-bogin-keenable]

Good catch — fixed. _call now builds plain result objects and JSON.stringifys the whole array, so the output is a valid JSON array as the tool description states (the previous .map(JSON.stringify).toString() emitted a comma-joined list).