How to see Gemini search queries
Gemini search queries can appear in search-call data from a Gemini API interaction with Google Search. The free tool shows the strings exposed in that run alongside source information at the scope the provider supplies.
Use this to research a topic with Gemini and inspect the current result. It is an API observation, not a private Gemini chat history or Google AI Mode query trace.
Run an observable Gemini searchChoose Gemini and inspect the query strings exposed in the current run before deciding whether to save it locally.Show Gemini queries- Open the tool with Gemini selected
Enter one short topic and set language and country if relevant. Native fanout uses a real API search; Search Ideas is a separate brainstorming mode.
- Read the current result
Inspect exposed query strings, search-action counts and source links. Zero exposed queries can be valid; a provider error is reported separately.
- Choose what to keep
Copy or export selected query strings for research. Optional local saves remove restricted Google Grounded Results and Search Suggestions; do not treat a saved run as a complete source archive.
Read Gemini strings in the correct response
The native tool currently uses Gemini Interactions with Google Search. In that documented contract, exposed strings appear in steps entries of type google_search_call, inside arguments.queries. This is a named API run, not access to private searches on gemini.google.com.
The field map below describes the contract and our original synthetic test example. It publishes no actual Gemini response. Citations in model_output annotations remain at run scope unless a finer relationship is supplied. A domain shown beside a query does not by itself establish a join.
If your response uses GenerateContent candidates, choose that separate contract. The checked Interactions parser does not import it as a fallback. A dedicated tutorial provides runnable parser code; this page explains what the displayed result means.
Continue with build the Interactions parser.
Interactions field map
Documentation and original synthetic structures, checked 10 October 2026; no actual grounding output.
| Field path | Read as | Boundary |
|---|---|---|
| steps[type=google_search_call].arguments.queries | Exposed strings | No complete private search history |
| model_output.content[].annotations | Run-level URL citations | No invented query join |
| google_search_result | Separate grounding payload | Not retained by the offline kit |
Live display and saved runs contain different data
On optional local save, the website removes Gemini grounding sources from sources and searchActions. Saved actions receive sourceScope not_stored; sourceEvidenceScope becomes not_stored_google_grounded_results. storageNotice explains that boundary in the export.
An empty source list in this saved format therefore establishes only that sources were not stored. It proves neither a source-free live run nor absence of provider citations. Label saved comparisons accordingly and do not add remembered sources.
A query selection is another, smaller contract: selected strings and run context without source fields. The offline validator distinguishes these forms and checks source stripping in saved Gemini full runs. Passing shape checks does not authorize a new reuse of actual data.
Continue with inspect the current export schema.
Retain observation, status and storage boundary together
Record exact input, country, language, model, time, visible strings and file type. Completed output with zero visible strings differs from a failed or partial attempt. The offline teaching example does not interpret absent status as success.
The current native mode checks status and native routing evidence through OpenRouter Responses, or through Chat Completions for Gemini. Only the optional direct Gemini adapter supplies a completed fallback when no provider status is provided. That fallback is not independent evidence of an extracted raw status. Stricter run checks must retain the information at the API boundary and review storage conditions.
Public example pages remain limited to previously reviewed OpenAI observations. The Gemini test kit uses only invented structures. This can suggest an editorial source-review or export task, but establishes no measured keyword volume, rank or general provider winner.
The boundaries are documented in the methodology.
Common questions
These short answers cover recurring questions from tool use and source review. They are not claimed search-volume data.
- Are these the searches from gemini.google.com?
- Not necessarily. The page describes query strings exposed by one Gemini API interaction with Google Search. It does not claim access to a private Gemini consumer-session trace.
- Why are some Gemini sources not saved?
- Google applies additional terms to Grounded Results and Search Suggestions. The tool removes restricted grounded data before an optional local save and does not publish it as a corpus.
Sources used
These sources support the functions and limits described here. They do not prove claims beyond that scope.
- Original runnable offline parser and edge-case test kitOpen source
Original educational parser code and 26 offline checks. Synthetic response fixtures are not provider output; exact strings, status states, source scope and retention exclusions are tested. No paid calls or new actual grounding data. The teaching helper explicitly differs from the live normalized wrapper.
- Current website export contracts and reference validationOpen source
Draft 2020-12 reference for the current native/modelled full runs, query selections and local comparisons. Schema checked against actual service/export functions with synthetic provider stubs and semantic negative tests. No root schemaVersion is invented for full runs. Legacy planner and handoff formats remain separate.
- Gemini Interactions: search calls and step typesOpen source
Supports Interactions steps, google_search_call arguments and output annotations. This is a different contract from GenerateContent candidates. Synthetic teaching fixtures do not establish publication rights for actual grounded output.
- AI Mode in Google Search: updates from Google I/O 2025Open source
Google states that AI Mode uses query fan-out to break a question into subtopics and issue multiple searches; it does not provide ai-fanout.com with a private AI Mode query trace.
- Gemini grounding with Google SearchOpen source
Documents Gemini 3.7 Flash Google Search, google_search_call arguments.queries, cited URL annotations and per-query billing.
- Gemini Developer API zero data retentionOpen source
Documents Interactions API store false and the unavoidable 30-day provider retention for prompts, context and output when Grounding with Google Search is used.
- Gemini API Additional TermsOpen source
Restricts caching, syndication, analysis and reuse of Google Grounded Results and Search Suggestions; drives the Gemini local-save redaction and public-example boundary.
- Direct OpenAI and Gemini native-fanout implementationOwner record
One direct provider request, strict provider allowlist, provider-exposed query/source parsing, honest zero-query output, CAPTCHA-first shared atomic quota and budget accounting, no raw server archive and separate public flag. Bounded live production fixtures passed for OpenAI and Gemini on 2026-08-26.