Claude Open Targets connectorWhat Claude can pull from Open Targets Platform.
The Claude Open Targets connector exposes 4 tools. All 4 read the platform, 0 write, 0 stay undocumented. Below: how Claude gets from a gene or disease name to an answer drawn from the platform, and where this connector's reach ends.
Verified Trustpilot reviews · AI, automation & growth agency
What changes when you are sizing up a drug target
Open Targets Platform supports finding and prioritising potential therapeutic targets by pulling public datasets together and scoring target-disease associations. The connector gives Claude guided access to that data: you ask in plain words, Claude resolves the names, builds the query and runs it on the platform.
Go from a common name to an ID. search_entities looks across targets, diseases, drugs, variants and studies and returns the ID and type of the top three matches for each term.
Learn what can be asked. get_open_targets_graphql_schema loads the relevant slice of the data structure, by category, before Claude writes its query. At least one category has to be named, and only that subset of the schema comes back.
Get the answer once, or across a list. query_open_targets_graphql runs one query against the platform API; batch_query_open_targets_graphql repeats the same query for a whole set of drugs, targets or diseases.
What Claude doesn't do here: change any data, the platform is for lookup. It works on the publicly available datasets the platform integrates. Nothing fires by itself either: no alert when the data changes. Scheduled monitoring belongs to an automation platform, covered in the Integrations hub.
The vocabulary, in one minute
Five words that help before you add Open Targets to Claude.
- Connector
- The link you add once between Claude and an outside source, so Claude can consult it while it answers you.
- Tool
- One named capability the connector opens to Claude. Claude picks which ones to call; the directory sheet names them.
- Authorization
- The service's screen where you grant the access Claude will use. Granted once per person, and withdrawable later.
- Approval
- The confirmation Claude waits for in the chat before any action that would change an account on your behalf.
- MCP
- The shared standard behind connectors: it lets an assistant such as Claude exchange data with an outside platform.
Add the Claude Open Targets connector in three steps
- 01
Find Open Targets
In Claude's settings, open Customize, then Connectors, and spot Open Targets in the list. On a Team or Enterprise workspace, an Owner or Primary Owner enables it for the whole organization first.
- 02
Start the connection
Click Connect on its row and follow the window that opens. If the link misbehaves one day, use Disconnect, then connect it again from the same list. Access can also be withdrawn from the service's side.
- 03
Read before you accept
Go through the authorization screen before confirming. The service writes it, not Claude, and it decides what the access covers. You can take the access back later the same way.
The 4 tools in the Claude Open Targets connector
Open Targets gives Claude 4 tools: 4 that read your account, 0 that change something in it.
One group only: all four tools consult the platform, often in this order, from name to ID to query. Names stay as Claude shows them.
- 4 read
What Claude reads (4)
4 toolsFour lookup tools, none of which writes to the platform.
get_open_targets_graphql_schema
Loads the description of the platform's data, filtered by category, so Claude knows which fields exist before it builds a query. At least one category is required.
search_entities
Turns everyday names into platform IDs: for each term, the top three matches across targets, diseases, drugs, variants and studies, each tagged with its type.
query_open_targets_graphql
Runs one query against the Open Targets Platform API and reports the result, once the IDs are resolved and the structure is known.
batch_query_open_targets_graphql
Repeats the same query with different values instead of firing calls one at a time. Built for comparing several drugs, targets or diseases at once.
What Claude asks before it acts
By default, Claude stops and asks for your go-ahead before every action it takes on an account for you.
For Open Targets, no source describes a confirmation specific to any of the four tools, which only consult data, so the general rule applies. Claude works with the rights of the connected person and nothing beyond them. On Team and Enterprise, workspace owners decide whether a member may let some actions through without a new prompt, and they can limit a connector for the entire organization, for example keeping reads and closing writes.
Which plans include it
None of the 819 sheets in the official directory shows availability by plan. For Open Targets, that answer isn't published anywhere.
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 enables the connector for the organization before members connect. For the current state, the Open Targets sheet in the official directory is the reference.
Where the Open Targets connector stops
A connector isn't an automation: Claude queries the platform while it is answering you, never in the background.
By default the sheet only exposes reading, an observed floor and not a guarantee: an admin can open actions that appear on no public sheet. The directory badge isn't a security audit, and Anthropic doesn't vouch for how a publisher's tools behave. Open Targets is the only source documenting this connector, through its tool descriptions. More connectors the team has studied are gathered in the Claude connectors hub.
Need help connecting Open Targets to Claude?
A person reads every message.