Claude Jedify connectorWhat Claude can do in your Jedify account.
The Claude Jedify connector lists 16 tools on its directory sheet. 13 read your data and its business definitions, 3 write by acting on a session or a question. Below: what Claude can ask, what your Jedify role unlocks beyond the sheet, and where the connector stops.
Verified Trustpilot reviews · AI, automation & growth agency
What changes when Claude goes through Jedify
Asking "what was churn last quarter" usually means finding the right table, the right definition of churn and someone who knows SQL. Jedify keeps a semantic layer, a map of your metrics, entities, business rules and terms, and the connector lets Claude ask its questions through that map. You get an answer built on your organization's own definitions, without needing to know where the data lives.
Get a quick figure. ask_a_single_question sends one question in plain words, and ask_questions runs several independent ones side by side. get_questions_data then returns the answer, the explanation, the SQL and the data for up to 5 inquiries at once.
Run a deeper investigation. ask_a_research_question launches a multi-step analysis that plans and executes many queries before handing back a report. wait_for_completion keeps an eye on it and returns the results when they are ready.
Ask the in-house expert. list_agents shows the custom agents your team built in Jedify, with their descriptions, and ask_an_agent puts a question to one of them. create_share_link produces a link to the full answer in the Jedify app for a colleague with a login in the same account.
What the sheet doesn't cover: the 16 tools are the question-asking set. Jedify says admins and data experts also get editing and building tools, decided by their role, and this page doesn't document those. SQL through run_sql_query is read-only. Nothing runs by itself either: Claude queries Jedify only when you ask in a conversation. A weekly figure pushed to a channel is automation work, a different job; the Integrations hub gathers that side.
The vocabulary in one minute
Five words that come up when you plug Jedify into Claude.
- Connector
- The link you set up once between Claude and an account you already hold, so Claude can work in it while it answers you.
- Tool
- A named action a connector opens to Claude. It decides which ones to call while it answers; the directory lists them by name.
- Authorization
- The service's own sign-in screen, where you hand Claude the access it will use. Granted once per person, and you can take it back.
- Approval
- The confirmation Claude waits for before completing something that changes your account, shown in the conversation at that point.
- MCP
- The common standard connectors are built on: it is what lets an assistant such as Claude talk to a service outside it.
Plug Jedify into Claude in three steps
- 01
Find Jedify in Claude
In Claude's settings, open Customize, then Connectors, and search for Jedify. On a Team or Enterprise workspace, an Owner or Primary Owner enables the connector first, then each member authenticates individually.
- 02
Sign in to Jedify
Click Connect on its row; your organization's Jedify login opens in a window of its own. Sign in there. If the link breaks later, Disconnect and connect again.
- 03
Check the authorization screen
Read what the screen asks before you accept. It belongs to Jedify, not Claude, and it sets the scope of the access. You can also withdraw that access later from the Jedify side.
The 16 tools, sorted by what they do
Jedify gives Claude 16 tools: 13 that read your account, 3 that change something in it.
Two groups: what Claude reads and what it acts on. The names stay in English, as Claude displays them.
- 13 read
- 3 write
What Claude reads (13)
13 toolsThirteen tools that ask questions, explore definitions and fetch results, without changing your data.
ask_a_research_question
Starts an autonomous, multi-step investigation: Jedify plans the work, runs many queries and comes back with a report rather than a single number.
ask_a_single_question
Puts one question in plain words to your data and opens a session for it automatically. When your account has several data sources, the source has to be specified.
ask_an_agent
Puts a question to one of the custom agents built in Jedify, which answers as the expert it was set up to be. You get session and inquiry IDs to follow like any other question.
ask_questions
Runs a batch of independent questions in parallel, each in its own session rather than one after the other. Their results are then collected like any other inquiry, through get_questions_data or wait_for_completion.
check_question_status
Tells whether a given inquiry has finished. For a research question, the research type has to be passed along with the ID.
contextualize
Loads your account's context: business model, glossary and the entities of every data source. Jedify presents it as the place to begin before any question.
data_collection_strategy
Lets Claude lay out a multi-step analysis plan before any query runs. Jedify reserves it for the question-asking mode, the one every Jedify role receives.
get_questions_data
Retrieves complete results for finished inquiries: the answer, its explanation, the SQL behind it and the data, for up to 5 inquiries per call.
get_semantic_layer_entities
Returns the full description of up to 10 entities from the semantic layer, in one call. With more than one data source, the taxonomy has to be named.
list_agents
Shows the custom agents of your Jedify account, with their descriptions and sample questions, and can filter them by a search term or cap how many come back.
run_sql_query
Runs a read-only SELECT directly against one data source. Jedify tells the assistant to use it rarely, since the semantic layer is the main way in.
search_semantic_layer
Finds entities by name, such as revenue or churn. For API sources, matches also come from descriptions and field names.
wait_for_completion
Waits on one or more inquiries and returns their results once done. After about three minutes it reports the work as still running, and Claude calls it again with the same IDs.
What Claude acts on (3)
3 toolsThree tools that create a link, stop a question or end the session. No source describes a confirmation of their own: the general rule below applies.
create_share_link
Approval: see the ruleProduces a link to the full answer of a session, narrative, charts and data, inside the Jedify app.
logout
Approval: see the ruleEnds the current Jedify session and clears its authentication.
stop_question
Approval: see the ruleHalts an inquiry that is still running, identified by its ID.
What Claude asks before acting
By default, Claude pauses and asks for your confirmation before each action it carries out on an account for you. The request appears in the conversation when it matters.
On Team and Enterprise, owners decide whether members can let some actions through without a new prompt. They can also cap what a connector does for the whole organization, for example keeping questions open and closing everything else; no one overrides that from their own account. Claude works with the access of the person connected and nothing more. On the Jedify side, your role decides which tools you even receive.
Which plans include it
Out of 819 sheets in the official directory, none displays availability per plan. That answer is published nowhere, connector by connector: it's a real gap in the catalogue.
The general rule is public: remote connectors are open to all users on Claude, Cowork, Claude Desktop and mobile. On Team and Enterprise, an Owner or Primary Owner turns the connector on for the organization before members connect. For the current state on your account, the Jedify sheet in the official directory is the reference.
Where this connector stops
A connector is not an automation. Claude calls these tools while it answers you; no question runs on a schedule or when your data changes.
The tool list is an observed floor, not a guarantee: an administrator can open actions that appear on no public sheet, and in Jedify's case your role adds some. The directory badge isn't a security audit, and Anthropic says on every sheet that it doesn't pick the tools a publisher exposes or vouch for their behavior. Only plug in publishers you trust. This connector is documented by Jedify alone, with no Claude help article.
Which Jedify tools does Claude actually get?
- Claude directory sheet ↗October 202616 tools, the question-asking set: questions, results, semantic layer and read-only SQL.
- Jedify MCP documentation ↗September 2026Tenant admins and data experts get three modes, including editing and building tools; users get the question-asking mode only.
What this page followsNo real contradiction: the sheet shows the shared base, the docs say your Jedify role adds editing and building tools on top. For a user role, the sheet is the full picture. For an admin or a data expert, the actual list is longer, and this page doesn't document it.
Need help connecting Jedify to Claude?
A person reads every message.