Deep Query - Stackone

Deep Query searches, filters, and aggregates the records Data Sync has stored, without calling the provider. It runs against StackOne’s index rather than a third-party API, so a query uses none of the provider’s rate limit and can answer questions the provider’s API has no parameter for. Deep Query reads the records Data Sync has stored so far, which incremental runs update in place. The live action stays available for the provider’s current state.

Why query instead of calling the provider

	Live action call	Deep Query
Where it reads from	The provider, on every call	StackOne’s index
Provider rate limits	Every call counts against them	No provider request, so nothing counts
Filtering	Only the query parameters that provider happens to expose	`term`, `terms`, `range`, `wildcard`, `match`, `match_phrase`, `exists`, and `bool` over any synced field up to four levels deep and outside lists of lists (what is indexed)
Full-text search	Rarely offered	Free-text across every text field, ranked by relevance
Aggregation	Not offered. Page through every record and count in your own code	One `aggs` query returns the numbers
Field selection	Whatever the provider chooses to return	`source` returns only the fields you name

A query is billed like an action call. Listing queryable fields is not.

Smaller responses, fewer tokens

The last two rows are where an agent gains the most. Asking for three fields instead of forty shrinks the response the model has to read by the same proportion. Asking for a headcount by department returns a dozen numbers from one aggregation, where a live call means paging through every employee record and tallying them inside the context window. Deep Query and Advanced Tool Search work on the same budget from opposite ends: one shrinks what tool responses cost, the other shrinks what tool definitions cost.

Which actions can sync

Deep Query covers any action with a sync turned on. An action can sync when its connector defines a sync for it. Actions that need parameters to run, such as fetching one record by ID, can’t. To see which actions can, open the connector in Connectors and select Library: actions that support Data Sync carry the sync icon. Turn sync on for the action in a Connector Profile first, as described in Data Sync.

Query from an agent

Once a sync has records, each synced action gains a companion tool named {action}_s1query, such as bamboohr_list_employees_s1query alongside bamboohr_list_employees. A list_s1query_queryable_fields tool tells the agent which fields it can filter and sort on. With Advanced Tool Search, search returns the _s1query entry just before the live action it copies, and the execute tool runs either one. Both tools stay available, so the agent chooses: the Deep Query tool to search or aggregate across everything cheaply, the live tool when it needs the provider’s current state. Records come back as the provider returned them to Data Sync, with the provider’s field names and nesting, which can differ from the live action’s result. Tool Policies apply to Deep Query reads the same way they apply to live action calls.

Try a query without code

Open Actions Request Tester from the account page and pick the (S1Query) entry for the action.

When to use it

Use Deep Query when the same data is read repeatedly and can be as old as the last sync run, or when an agent needs to search or aggregate across the whole record set rather than page through it. For data that has to reflect the provider right now, call the action directly. Both tools stay available to the agent. To be told when a synced record changes rather than querying for it, turn on Data Sync events.

Data Sync

Turn on the sync that Deep Query reads from, and set how often it refreshes.

Query from your own code

For developers querying the index from their own backend, with filters and aggregations.

​Why query instead of calling the provider

​Smaller responses, fewer tokens

​Which actions can sync

​Query from an agent

​Try a query without code

​When to use it