Asker Tools - Jedify

Your agent calls these tools to answer questions from your data. Editor and Builder modes include all of them too — except data_collection_strategy, which is Asker-only — plus their own tools.

Context

Tool	Purpose
`contextualize()`	Load the account’s context — business model, glossary, and the entities of every data source. Start here.
`data_collection_strategy(strategy)`	Plan a multi-step analysis before querying. Asker mode only.

Ask

Tool	Purpose
`ask_a_single_question(question, data_source_id?)`	Ask one question; a session is created automatically. `data_source_id` is required when the account has more than one data source.
`ask_questions(questions)`	Ask several independent questions in parallel, each in its own session.
`ask_a_research_question(question)`	Run a multi-step autonomous investigation that plans and executes many queries and writes a report.
`stop_question(inquiry_id)`	Stop a running inquiry.

Custom agents

The experts your team has built in Jedify — each with its own persona, instructions, and entity scope.

Tool	Purpose
`list_agents(search?, limit?)`	List the account’s custom agents with their descriptions and sample questions.
`ask_an_agent(agent_id, question, session_id?)`	Ask a question of one agent; it answers as the expert it was built to be. Returns a `session_id` and `inquiry_id` to poll like any other question.

Get results

Tool	Purpose
`check_question_status(inquiry_id)`	Check whether an inquiry has finished. Pass `inquiry_type: "research"` for research questions.
`wait_for_completion(inquiry_ids)`	Wait for inquiries to finish, then return their results. After about three minutes it returns `in_progress`; your agent calls it again with the same ids to keep waiting.
`get_questions_data(inquiry_ids)`	Retrieve full results — answer, explanation, SQL, and data — for up to 5 inquiries at once.
`create_share_link(session_id)`	Create a link that opens the full answer — narrative, charts, and data — in the Jedify app. Opening it requires a Jedify login in the same account; it is not a public link.

Explore the semantic layer

Tool	Purpose
`search_semantic_layer(search_term)`	Find entities by name (e.g. “revenue”, “churn”). Entities from API sources also match on their descriptions and field names.
`get_semantic_layer_entities(node_ids, taxonomy_id?)`	Get full details for up to 10 entities. `taxonomy_id` is required when the account has more than one data source.

Advanced

Tool	Purpose
`run_sql_query(query, data_source_id?)`	Run a read-only SQL `SELECT` directly against a data source. Your agent is instructed to use this rarely — the semantic layer is the primary interface.
`call_api(data_source_id, node, operation, params?)`	Run one declared read operation against a connected API source (such as HubSpot) and return the raw response. Listed only when your account has a connected API source. Questions about API data normally go through `ask_a_single_question` with the source’s `data_source_id` instead.
`logout()`	End the session and clear authentication.