Export fanout queries as CSV or JSON

Select the queries you need and use Keyword CSV or Selected JSON in the result. Download CSV and Download JSON preserve a full current run. Saved runs can be exported together from local history; these three export scopes contain different fields.

Use this when reviewing queries in a spreadsheet, transferring them to a keyword list or documenting a dated comparison.

Start with a current resultRun one short topic, then select only the queries you want to investigate further.Open the tool and local history

Choose the appropriate export scope

When a result is available, select individual query rows. Keyword CSV contains only that selection, with the input, provider, model, timestamp, locale and evidence status. Selected JSON structures the same information. Sources are not attached to the query selection.

Download JSON preserves the full currently visible result contract. Download CSV creates a flat table with separate query, search-action-source and run-source rows. Row types and source scope matter: a source shared by a search action is not automatically evidence for one individual query.

For source scope, follow the AI citation and source guide.

For several saved runs, use the checkboxes under Saved on this device, then JSON or CSV. That export documents a local comparison. It is not an independent provider benchmark.

Selection, current run or comparison?

The controls serve different tasks; keep the appropriate scope with your working notes.

You need	Export	Important limit
A curated keyword list	Keyword CSV or Selected JSON	Selected queries only; no query-to-source mapping.
The current result contract	Download JSON or Download CSV	Keep API fields separate from editorial interpretation.
Several saved runs	JSON or CSV in local history	Gemini grounded sources are absent after the save filter.

Read the keyword CSV correctly

The header below comes from the query-selection exporter. query contains the selected string; query_index preserves its original position. observed_at carries the timestamp stored with the run. A later export time does not turn it into a new observation.

evidence_status keeps modelled and observed results distinct. intent and reason are empty for native string queries; modelled search ideas can carry these fields from the generated result. Empty fields are not lost keywords and should not be filled with an invented measured intent.

Import CSV as comma-separated UTF-8 text and treat search phrases as text. Check the row count, special characters, provider and date before transferring the selection into another tool. Acquire search volume and competition in a separate keyword dataset.

For the eventual page decision, use the query fanout SEO workflow.

Query-selection export header

keyword,provider,model,observed_at,country,language,evidence_status,query_index,query,intent,reason

Account for local history, expiry and Gemini sources

Save on this device is an explicit action. History lives in IndexedDB for this browser and exact origin. It holds at most 20 results for up to 30 days. Another browser, device or Vercel preview does not share that history.

Export important work before expiry or clearing website data. The persistent-storage request is not a guarantee: the browser decides, and the application's 30-day limit still applies. There is no account sync or recovery from a server archive.

Before saving a Gemini result locally, the tool removes Google grounded sources and search suggestions. A later comparison export therefore has no saved Gemini sources. Their absence does not prove the original run had no sources. Follow the applicable terms when sharing current results.

For interpreting these differences, read the local model-comparison guide.

Verified roundtrip: produce a selection and read it independently

The three exercise files were produced by this website's actual selection-export functions. The input run is entirely synthetic and labelled accordingly; no provider was called. Original positions one and three were selected from three strings. JSON preserves those indices rather than renumbering after selection.

The first selected string is Café, query examples. The second contains quotes and a newline. The CSV therefore has more physical lines than records. An independent CSV reader recovered two records, eleven columns, unchanged strings, run metadata and original indices one and three. The JSON files were also checked against the public reference schema.

intent and reason remain empty for native string queries. The selection contains neither sources nor providerResponseStatus. Interpreting completed, incomplete or error states also requires the full run. Even though a teaching field carries the ordinary native evidence label, this remains a synthetic structural test rather than an actual search observation.

For the separate field families, read the fanout result-schema guide.

Download the full run, selection JSON, selection CSV and verification notes together. This file roundtrip checks the exporter and independent readback; it does not establish an executed Excel or Google Sheets import or an operating-system download-dialog save. Recheck strings and text types in your actual receiving application.

Check an actual selection and its receiver in the two-query handoff exercise.

Import the selection export with the tested Google Sheets import.

Common questions

These short answers cover recurring questions from tool use and source review. They are not claimed search-volume data.

Can I recover a saved result after 30 days?: Local history expires after no more than 30 days. A previously downloaded export is your own working record; ai-fanout.com does not maintain a raw-result recovery archive.
Why does the keyword CSV contain no sources?: It exports selected query strings and run metadata. Adding a general source list could imply an unsupported relationship to those individual queries.

Sources used

These sources support the functions and limits described here. They do not prove claims beyond that scope.

Original annotated workflow, controls, prompts and export exercises
Original diagrams and editorial annotations; only the previously approved normalized OpenAI records. Synthetic selection files use actual producers and independent CSV readback, not new provider output or Excel/Sheets import. Own dated public-page checklist makes indexing and outcomes explicitly unproven.
Open source
Current website export contracts and reference validation
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.
Open source
ai-fanout.com input, selection export and local-history implementation
The input controls, separate native/modelled modes, three export scopes, query-selection CSV fields, and browser-local history are described from the site's actual implementation. The planning prompt and input examples are original editorial instructions, not provider output or measured keyword demand.
Open source
MDN: IndexedDB API
Explains browser-side structured storage and its origin boundary. The site's twenty-entry and thirty-day limits come from its own implementation, not from a universal browser quota.
Open source
Opt-in browser-local result history
Explicit save action, IndexedDB-only storage, maximum 20 entries, 30-day expiry, individual deletion, clear-all and local JSON/CSV export without accounts or shared response caching.
Owner record