n8n

Latest Audit Result

The n8n-integration (v0.3.3) uses n8n declarative REST routing against api.exa.ai — no exa-js SDK dependency. It exposes Search (auto/fast/instant + deep variants), Contents, and Answer. The README over-declares Find Similar and Research operations that are absent from the node code. /research is now deprecated (sunset May 1 2026), so its absence is correct. /findSimilar is also deprecated per skill docs. Multiple newer API parameters are missing from existing operations (livecrawl 'preferred', maxAgeHours, structured summary/highlights, answer text/stream flags). The node also sends a 'deep-max' type that is not in the current API enum.

Missing Features (17)

• README declares Find Similar and Research resources/operations that are NOT implemented in the node TypeScript code (resource selector only has search, contents, answer)
• Deep search exposes 'deep-max' type which is NOT in the current /search API enum (neural, fast, auto, deep-lite, deep, deep-reasoning, instant) — may produce 400 errors
• Regular search type options are missing 'neural' which is a valid API enum value
• Livecrawl options list is missing 'preferred' (added June 2025) — only offers always/never/fallback
• Contents/search: 'maxAgeHours' freshness control parameter is not exposed (the normative content freshness mechanism per Feb 2026 changelog)
• Answer endpoint: missing 'text' boolean parameter (controls inclusion of full text in search results)
• Answer endpoint: missing 'stream' boolean parameter (SSE streaming)
• README mentions 'keyword' as a search type but the node does not expose it and the API no longer lists it in the enum — README is inaccurate
• Livecrawl: add 'preferred' option (tries fresh, falls back to cache) alongside existing always/never/fallback
• Contents options: upgrade 'summary' from boolean-only to support {query, schema} for structured summaries
• Contents options: upgrade 'highlights' to accept {query} for guided highlight extraction (maxCharacters already supported)
• Contents options: add 'maxAgeHours' numeric parameter for fine-grained content freshness (replaces livecrawl enum for new integrations)
• Contents options: add 'subpageTarget' string/array field for focused subpage crawling (e.g. docs, about)
• Answer: expose 'text' boolean and 'stream' boolean parameters already in the API spec
• Search: add 'neural' to the regular search type dropdown
• Deep search: remove or validate 'deep-max' against the actual API — it's not in the current OpenAPI spec
• README: remove references to Find Similar and Research operations that are not implemented, and remove 'keyword' search type claim