Update Exa integration examples to use search()

[10ishq]

Updates the Exa integration examples:

• Switched from search_and_contents() to search() in both examples
• Example 1 now shows basic search with title/URL display
• Example 2 builds OpenAI summarization context from search result titles and URLs
• Removed references to text and highlights content options

Greptile Summary

This PR updates both the MDX integration page and the companion Jupyter notebook to use Exa's search() method instead of search_and_contents(), removing page-text retrieval from the examples.

• Example 1 is updated correctly: exa.search() returns title and URL per result, and the display loop now prints those two fields with a blank separator line.
• Example 2 retains the "summarize search results" narrative but now passes only titles and URLs as the OpenAI context, leaving the model with no actual page content to summarize — the RAG pipeline is effectively hollow and will produce misleading or fabricated outputs for anyone who runs this example.
• The notebook was also significantly reformatted (indentation, removal of execution_count/outputs fields), which is cosmetic but makes diffs harder to read in future PRs.

Confidence Score: 3/5

Example 2 demonstrates a broken RAG pattern — the LLM receives only link titles, not page content — so merging as-is would publish misleading documentation to users.

The summarization example now feeds the OpenAI model a list of page titles and URLs while still asking it to summarize search results. Any user who follows this example will get a low-quality or fabricated summary because there is no actual text content in the context. This is the core intent of the integration guide and it is demonstrably undermined by the change.

Both content/integrations/other/exa.mdx and cookbook/integration_exa.ipynb share the same broken summarization context in Example 2 and need to be addressed together.

Sequence Diagram

sequenceDiagram
    participant User
    participant search_and_summarize
    participant search_with_exa
    participant Exa
    participant OpenAI

    User->>search_and_summarize: query
    search_and_summarize->>search_with_exa: "query, num_results=5"
    search_with_exa->>Exa: exa.search(query, num_results)
    Note over Exa: Returns title + URL only<br/>(no page text)
    Exa-->>search_with_exa: SearchResponse (title, url, id)
    search_with_exa-->>search_and_summarize: results
    search_and_summarize->>search_and_summarize: "build context = title + URL only"
    search_and_summarize->>OpenAI: Summarize... + context (titles/URLs)
    Note over OpenAI: No actual page content to summarize
    OpenAI-->>search_and_summarize: low-value summary
    search_and_summarize-->>User: print summary

Loading

Prompt To Fix All With AI

Fix the following 1 code review issue. Work through them one at a time, proposing concise fixes.

---

### Issue 1 of 1
content/integrations/other/exa.mdx:111-122
**Summarization context is titles/URLs only — no content to summarize**

After switching to `exa.search()`, the context built for OpenAI contains only `r.title` and `r.url` — no page text. The system prompt then instructs GPT to "Summarize the following search results clearly and concisely," but there is no substantive content to summarize; the model only sees a list of link titles. The output will be low-quality or hallucinatory, which misrepresents what a real Exa+LLM RAG pipeline would look like. To get actual page content you need `exa.search_and_contents()` with `text=True` or `highlights=True`, or the prompt/description should be updated to frame this as link recommendation rather than content summarization.

_{Reviews (1): Last reviewed commit: "use search instead of search_and_content..." | Re-trigger Greptile}

[vercel]

@10ishq is attempting to deploy a commit to the langfuse Team on Vercel.

A member of the Team first needs to authorize it.

[claude]

Claude Code Review

This pull request is from a fork — automated review is disabled. A repository maintainer can comment @claude review to run a one-time review.

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[CLAassistant]

All committers have signed the CLA.

[10ishq]

Hey @marcklingen — updates the Exa integration examples to current API conventions. CLA is signed. Would appreciate a review!

[marcklingen]

thank you! @jannikmaierhoefer can you review this?