Loading ToolBench
Tools that enable LLMs to interact directly with the weaviate API.
Activate a deactivated database user account. Use this tool to activate a user with a `db` user type in the database. It should be called when re-enabling user access is needed.
Add new permissions to a role without affecting existing ones. Use this tool to assign additional permissions to a specified role within a system, ensuring that current permissions remain unchanged. Note: Understanding the request schema is necessary to properly create the stringified JSON input object for execution. This operation also requires path parameters. Modes: - GET_REQUEST_SCHEMA: Returns the schema. Only call if you don't already have it. Do NOT call repeatedly if you already received the schema. - EXECUTE: Performs the operation with the provided request body JSON. Note: You must also provide the required path parameters when executing. If you need the schema, call with mode='get_request_schema' ONCE, then execute.
Add a new property to an existing collection schema. Use this tool to add a new property definition to a specified collection (className) within the schema. It is called when a new property needs to be incorporated into an existing data structure. Note: Understanding the request schema is necessary to properly create the stringified JSON input object for execution. This operation also requires path parameters. Modes: - GET_REQUEST_SCHEMA: Returns the schema. Only call if you don't already have it. Do NOT call repeatedly if you already received the schema. - EXECUTE: Performs the operation with the provided request body JSON. Note: You must also provide the required path parameters when executing. If you need the schema, call with mode='get_request_schema' ONCE, then execute.
Add a reference to an object's property. This tool adds a new reference to a specified property of a data object. The object is identified by its class and UUID. Use this to link related objects within a dataset. Note: Understanding the request schema is necessary to properly create the stringified JSON input object for execution. This operation also requires path, query parameters. Modes: - GET_REQUEST_SCHEMA: Returns the schema. Only call if you don't already have it. Do NOT call repeatedly if you already received the schema. - EXECUTE: Performs the operation with the provided request body JSON. Note: You must also provide the required path, query parameters when executing. If you need the schema, call with mode='get_request_schema' ONCE, then execute.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-03-17 | A | 82 | - | v1 |
Assign roles to a specific group. Use this tool to assign one or more roles to a designated group, identified by its ID.
Assign roles to a user in the system. Use this tool to assign one or more roles to a specific user, allowing flexibility in user permissions and access control.
Batch create cross-references between items in a collection. Use this tool to efficiently establish multiple cross-references between items in a collection at once. Ideal for handling large datasets where establishing individual links manually would be inefficient. Note: Understanding the request schema is necessary to properly create the stringified JSON input object for execution. This operation also requires query parameters. Modes: - GET_REQUEST_SCHEMA: Returns the schema. Only call if you don't already have it. Do NOT call repeatedly if you already received the schema. - EXECUTE: Performs the operation with the provided request body JSON. Note: You must also provide the required query parameters when executing. If you need the schema, call with mode='get_request_schema' ONCE, then execute.
Register multiple data objects in a single request. This tool registers multiple data objects in one request for efficiency. It validates metadata and schema values for each object. The operation is idempotent, meaning if an object with a given UUID already exists, it will be overwritten. Note: Understanding the request schema is necessary to properly create the stringified JSON input object for execution. This operation also requires query parameters. Modes: - GET_REQUEST_SCHEMA: Returns the schema. Only call if you don't already have it. Do NOT call repeatedly if you already received the schema. - EXECUTE: Performs the operation with the provided request body JSON. Note: You must also provide the required query parameters when executing. If you need the schema, call with mode='get_request_schema' ONCE, then execute.
Cancel a backup by its ID from a specified backend storage. Use this tool to delete a backup using its unique ID from a specified backend storage location, ensuring the backup is no longer retrievable.
Cancel an active replication operation by ID. Use this tool to request the cancellation of an ongoing replication operation specified by its ID. The operation is stopped and marked as 'CANCELLED' but is not automatically deleted.
Cancel an active replication operation. Use this tool to remove and cancel a specific replication operation. It ensures that active processes are halted and resources are cleaned up before deletion.
Check the status of a backup restoration process. Use this tool to manually check the status of a specific backup restoration on a given backend, using the backup ID. Ideal when automatic polling is disabled.
Get the current status of a backup creation process. Use this tool to manually check the status of a backup creation process by its ID on the specified backend. Ideal for when waiting for completion is disabled or not suitable.
Check if a role has specific permissions in the system. Use this tool to verify whether a particular role possesses certain permissions within the system. It should be called when there's a need to confirm role access rights. Note: Understanding the request schema is necessary to properly create the stringified JSON input object for execution. This operation also requires path parameters. Modes: - GET_REQUEST_SCHEMA: Returns the schema. Only call if you don't already have it. Do NOT call repeatedly if you already received the schema. - EXECUTE: Performs the operation with the provided request body JSON. Note: You must also provide the required path parameters when executing. If you need the schema, call with mode='get_request_schema' ONCE, then execute.
Check if the Weaviate instance is running properly. Use this tool to perform a basic health check on a Weaviate instance, ensuring it is running and responding to HTTP requests. Useful for monitoring and maintenance tasks.
Check if the Weaviate instance is ready to accept traffic. Use this tool to determine if the Weaviate instance has completed its startup routines and is ready for operations such as data import and queries. Ideal for readiness checks in container orchestration contexts like Kubernetes.
Create a new alias for a collection in Weaviate. This tool creates a new alias mapping between an alias name and a collection (class) in Weaviate. Use it to set up alternative names for accessing collections.
Initiates backup creation for specified collections. This tool starts the process of creating a backup for the specified collections on a designated backend storage using Weaviate. The backup is compressed using gzip by default, and the system remains operational during the backup process. Use this tool when you need to ensure data is safely backed up without interrupting service.
Create a new database user and obtain an API key. This tool is used to create a new database user by specifying a name. It returns an API key for the created user.
Create a new data object in Weaviate. This tool creates a new data object in Weaviate, ensuring that the object's metadata and schema values are validated. It's useful for adding new data objects when there is no existing object with the same ID. Note: Understanding the request schema is necessary to properly create the stringified JSON input object for execution. This operation also requires query parameters. Modes: - GET_REQUEST_SCHEMA: Returns the schema. Only call if you don't already have it. Do NOT call repeatedly if you already received the schema. - EXECUTE: Performs the operation with the provided request body JSON. Note: You must also provide the required query parameters when executing. If you need the schema, call with mode='get_request_schema' ONCE, then execute.
Create a new role with specified permissions. Use this tool to create a new role in the system and define its permissions. It should be called when there is a need to set up access controls for new role-based functionalities. Note: Understanding the request schema is necessary to properly create the stringified JSON input object for execution. Modes: - GET_REQUEST_SCHEMA: Returns the schema. Only call if you don't already have it. Do NOT call repeatedly if you already received the schema. - EXECUTE: Performs the operation with the provided request body JSON. If you need the schema, call with mode='get_request_schema' ONCE, then execute.
Create a new collection (class) in Weaviate. This tool is used to define and create a new collection (class) in Weaviate. It provides explicit control over the schema definition, which is essential when `AutoSchema` is not relied upon. Use this tool to ensure accurate schema setup in a Weaviate instance. Note: Understanding the request schema is necessary to properly create the stringified JSON input object for execution. Modes: - GET_REQUEST_SCHEMA: Returns the schema. Only call if you don't already have it. Do NOT call repeatedly if you already received the schema. - EXECUTE: Performs the operation with the provided request body JSON. If you need the schema, call with mode='get_request_schema' ONCE, then execute.
Create new tenants in a specified collection. Use this tool to create one or more new tenants for a specified collection when multi-tenancy is enabled. Useful for managing tenants within a Weaviate database. Note: Understanding the request schema is necessary to properly create the stringified JSON input object for execution. This operation also requires path parameters. Modes: - GET_REQUEST_SCHEMA: Returns the schema. Only call if you don't already have it. Do NOT call repeatedly if you already received the schema. - EXECUTE: Performs the operation with the provided request body JSON. Note: You must also provide the required path parameters when executing. If you need the schema, call with mode='get_request_schema' ONCE, then execute.
Deactivate a database user account. Use this tool to deactivate a user with the type 'db' in the database. Provide the user ID to deactivate their account and prevent access to the database.
Delete an existing alias from the system. This tool removes an alias from the system without affecting the underlying collection. It should be called when an alias is no longer needed.
Schedule deletion of all replication operations across the system. This tool is used to schedule the deletion of all replication operations across all collections, shards, and nodes. It should be called when you need to clear all replication tasks in the system.
Delete a data object from a specified collection using its UUID. Use this tool to remove a specific data object from a Weaviate collection, identified by its `className` and `id` (UUID). Useful for managing data within a Weaviate instance.
Delete a specific database user. This tool deletes a specified user from the database. It should be called when there is a need to remove a user, except the current active user.
Deletes multiple data objects using specified filter criteria. This tool removes data objects based on a filter. Use it when you need to delete multiple objects that match specific criteria within the limit. The deletion is performed in the order objects match the filter. If you need to delete beyond the limit, repeat the process until finished. Note: Understanding the request schema is necessary to properly create the stringified JSON input object for execution. This operation also requires query parameters. Modes: - GET_REQUEST_SCHEMA: Returns the schema. Only call if you don't already have it. Do NOT call repeatedly if you already received the schema. - EXECUTE: Performs the operation with the provided request body JSON. Note: You must also provide the required query parameters when executing. If you need the schema, call with mode='get_request_schema' ONCE, then execute.
Delete a specific reference from an object's property. Use this tool to remove a reference from a specific property of a data object in a collection, identified by its class name and UUID. Call this tool when you need to delete a reference from an object's property. Note: Understanding the request schema is necessary to properly create the stringified JSON input object for execution. This operation also requires path, query parameters. Modes: - GET_REQUEST_SCHEMA: Returns the schema. Only call if you don't already have it. Do NOT call repeatedly if you already received the schema. - EXECUTE: Performs the operation with the provided request body JSON. Note: You must also provide the required path, query parameters when executing. If you need the schema, call with mode='get_request_schema' ONCE, then execute.
Delete a role and revoke its permissions system-wide. Use this tool to delete a role from the system, removing it and revoking the associated permissions from all users who had it.
Permanently delete a collection from the schema. Call this tool to remove a collection definition and all its data objects from the schema permanently.
Permanently delete specified tenants from a collection. Use this tool to delete one or more tenants from a specified collection in Weaviate. This action is irreversible and will remove all data related to the specified tenants.
Retrieve links to available REST API endpoints. This tool fetches links to other endpoints within the REST API, aiding in the discovery and navigation of the available API options.
Execute multiple GraphQL queries in a single request. Executes multiple GraphQL queries sent in a single network request for efficiency. Ideal for performing batch queries in Weaviate. Note: Understanding the request schema is necessary to properly create the stringified JSON input object for execution. Modes: - GET_REQUEST_SCHEMA: Returns the schema. Only call if you don't already have it. Do NOT call repeatedly if you already received the schema. - EXECUTE: Performs the operation with the provided request body JSON. If you need the schema, call with mode='get_request_schema' ONCE, then execute.
Retrieve the status of a specific replication operation. Use this tool to get current status and detailed information about a replication operation by its unique ID. It can optionally include historical progress data of the operation.
Fetch role details using its name. This tool retrieves the details of a specific role by its name from the Weaviate service. It should be called when information about a role is required, typically for authorization or configuration purposes.
Fetch the current sharding state and replica details for collections. This tool retrieves the current sharding state for all collections or a specified collection, including replica locations and statuses. If needed, it can also provide the state for a specific shard within a collection.
Forcefully delete replication operations with caution. This tool is used to forcefully delete operations from the FSM in Weaviate. It should be called when there is a need to remove replication operations without performing state checks, which may lead to data corruption or loss. Ensure replication engine workers are scaled to 0 before using this to prevent in-flight operations.
Retrieve details about the authenticated user and their roles. This tool retrieves information about the current user, including their username and assigned roles, from the authenticated session.
Retrieve status and results of a classification task. Use this tool to obtain the status, metadata, and results of an ongoing or completed classification task by providing its unique ID.
Retrieve status information about all nodes in a cluster. Fetches status details for all nodes in the cluster, with adjustable detail level using the output parameter.
Retrieves status of shards for a specified collection. Use this tool to get the status of all shards associated with a specific collection (`className`) in Weaviate. For multi-tenant collections, specify the `tenant` query parameter to retrieve shard status for a particular tenant.
Retrieve tenants for a specified collection. This tool retrieves a list of all tenants associated with a given collection. Call this tool to obtain tenant information for a specified collection in the Weaviate service.
Retrieve a data object using collection name and UUID.
Retrieve information about a specific database user. Use this tool to get detailed information about a database user, including their roles, status, and type. Call this tool when you need to know specific details about a database user's profile or attributes.
Retrieve groups assigned to a specific role. Use this tool to get a list of all groups that have been assigned a particular role, identified by the role's name.
Fetches OIDC discovery details for Weaviate authentication. This tool retrieves OpenID Connect (OIDC) discovery information for Weaviate if OIDC authentication is configured. It provides essential details like the token issuer URL, client ID, and required scopes.
Retrieve all roles and their assigned permissions. Use this tool to obtain detailed information on all roles and the specific permissions associated with each role.
Retrieve roles assigned to a specific group. Fetches all roles associated with a group identified by its name and type (options: `db` or `oidc`).
Retrieve status of nodes hosting shards for a collection. This tool fetches status information for nodes that host shards of a specified collection. Use it to monitor or diagnose the nodes in a distributed database cluster, customizing the level of detail with the `output` query parameter.
Retrieve details about a specific tenant's status. Call this tool to get information about a specific tenant within a specified collection, including their current activity status.
Retrieve all roles assigned to a specific user. Use this tool to get a list of all roles for a specified user, identified by their ID and user type (`db` or `oidc`).
Retrieve users with a specific role assignment. Use to obtain a list of users assigned to a particular role by role ID.
Get Weaviate cluster Raft protocol statistics. Use this tool to retrieve detailed statistics about the internal Raft consensus protocol state for a Weaviate cluster. This can help monitor the cluster's status and performance.
Get meta-information about a Weaviate instance. Use this tool to obtain details such as version, modules, and network hostname of a running Weaviate instance. Useful for monitoring, compatibility checks, or communication between instances.
Initiate a classification task in the background. Use this tool to start a background classification task with specific parameters. This will initiate the process, and you can monitor its status and retrieve results using the GET /classifications/{id} endpoint. Note: Understanding the request schema is necessary to properly create the stringified JSON input object for execution. Modes: - GET_REQUEST_SCHEMA: Returns the schema. Only call if you don't already have it. Do NOT call repeatedly if you already received the schema. - EXECUTE: Performs the operation with the provided request body JSON. If you need the schema, call with mode='get_request_schema' ONCE, then execute.
Retrieve all backup IDs and their statuses. Use this tool to get a list of all created backup IDs along with their current statuses. Useful for monitoring or managing backup data.
Retrieve a list of data objects from a specified collection. Use this tool to get a list of data objects from a particular collection by providing the collection name parameter. The objects are returned in reverse order of creation.
Retrieve all database users and their roles and statuses. Use this tool to obtain a comprehensive list of all users within the database, along with detailed information about their roles and current status. This is useful for auditing and managing user permissions in the database.
Retrieve all distributed tasks in the cluster. Use this tool to get an overview of all distributed tasks currently active or available within the cluster.
Retrieve registered replication operations and details. Call this tool to get information about current replication operations in Weaviate, filtered by collection, shard, or node ID if needed.
Replace existing references for an object in Weaviate. Use this tool to replace all current references of a specific property for an object in Weaviate. Specify the object's collection name, UUID, and the reference property to update. Useful for updating linked data within a dataset. Note: Understanding the request schema is necessary to properly create the stringified JSON input object for execution. This operation also requires path, query parameters. Modes: - GET_REQUEST_SCHEMA: Returns the schema. Only call if you don't already have it. Do NOT call repeatedly if you already received the schema. - EXECUTE: Performs the operation with the provided request body JSON. Note: You must also provide the required path, query parameters when executing. If you need the schema, call with mode='get_request_schema' ONCE, then execute.
Initiates replication of a shard replica to a target node. This tool starts an asynchronous process to move or copy a specific shard replica from its current node to a specified target node. It handles data copying, synchronization, and may decommission the source replica.
Restore collections from a specified backup. Initiates the restoration of collections from a backup stored on a designated backend. Ensure the target cluster has identical node configurations as the source cluster, and that the collections do not already exist on the target. Note: Understanding the request schema is necessary to properly create the stringified JSON input object for execution. This operation also requires path parameters. Modes: - GET_REQUEST_SCHEMA: Returns the schema. Only call if you don't already have it. Do NOT call repeatedly if you already received the schema. - EXECUTE: Performs the operation with the provided request body JSON. Note: You must also provide the required path parameters when executing. If you need the schema, call with mode='get_request_schema' ONCE, then execute.
Retrieve details about a specific alias by its name. Use this tool to obtain information about an alias, such as the collection (class) it points to, by specifying the alias name.
Retrieve all aliases from the system. Fetches a list of all aliases, with an option to filter by collection (class) name for specific aliases.
Retrieve the schema of a specified collection. Fetches the definition of a specific collection by `className`, detailing its properties, configuration, and vectorizer settings.
Retrieve definitions of all classes in the database schema. Use this tool to obtain the definitions of all collections (classes) currently present in the database schema. It provides a comprehensive overview of the database structure.
Retrieve available group names for a specified type. This tool retrieves a list of all available group names for a specified group type, either 'oidc' or 'db'. Use it when you need to list groups of a certain type.
Revoke roles from a specified group to manage permissions. This tool allows the revocation of roles from a specified group, which is useful for managing and updating group permissions within a system using Weaviate.
Revoke permissions from a specified role. This tool is used to revoke permissions from a specified role in the Weaviate system. If all permissions are removed, the role will be deleted. Note: Understanding the request schema is necessary to properly create the stringified JSON input object for execution. This operation also requires path parameters. Modes: - GET_REQUEST_SCHEMA: Returns the schema. Only call if you don't already have it. Do NOT call repeatedly if you already received the schema. - EXECUTE: Performs the operation with the provided request body JSON. Note: You must also provide the required path parameters when executing. If you need the schema, call with mode='get_request_schema' ONCE, then execute.
Remove roles from a specified user in the system. Use this tool to revoke one or more roles assigned to a user by specifying their user ID. It's useful for updating user permissions in the system.
Revoke and regenerate the API key for a database user. This tool is used to revoke the current API key for a specified database user and generate a new one. It should be called when an API key needs refreshing or if security has been compromised.
Redirect an alias to a different collection. Use this tool to update an existing alias so it points to a new collection. This allows alias redirection without changing its name.
Update settings of an existing collection. Use this tool to update the configuration settings of an existing collection by specifying the className and the new settings. This tool is useful when you need to change mutable settings of a collection but not its name or properties. Note: Understanding the request schema is necessary to properly create the stringified JSON input object for execution. This operation also requires path parameters. Modes: - GET_REQUEST_SCHEMA: Returns the schema. Only call if you don't already have it. Do NOT call repeatedly if you already received the schema. - EXECUTE: Performs the operation with the provided request body JSON. Note: You must also provide the required path parameters when executing. If you need the schema, call with mode='get_request_schema' ONCE, then execute.
Update specific properties of a data object. This tool updates properties of an existing object in a specified collection using JSON merge patch semantics. The object is identified by its collection name and UUID. Use this to modify selected fields of a data object while ensuring validation and timestamp updates. Note: Understanding the request schema is necessary to properly create the stringified JSON input object for execution. This operation also requires path, query parameters. Modes: - GET_REQUEST_SCHEMA: Returns the schema. Only call if you don't already have it. Do NOT call repeatedly if you already received the schema. - EXECUTE: Performs the operation with the provided request body JSON. Note: You must also provide the required path, query parameters when executing. If you need the schema, call with mode='get_request_schema' ONCE, then execute.
Replace properties of a data object using its class and ID. This tool updates the properties of an existing object identified by its class name and UUID. The request must include the complete object definition with new values to replace the existing ones. Note: Understanding the request schema is necessary to properly create the stringified JSON input object for execution. This operation also requires path, query parameters. Modes: - GET_REQUEST_SCHEMA: Returns the schema. Only call if you don't already have it. Do NOT call repeatedly if you already received the schema. - EXECUTE: Performs the operation with the provided request body JSON. Note: You must also provide the required path, query parameters when executing. If you need the schema, call with mode='get_request_schema' ONCE, then execute.
Update the status of a specific shard in a collection. Use this tool to update the status of a specific shard within a Weaviate collection. It is typically used to change the shard's status to `READY` or `READONLY` after addressing issues that affected its operation.
Update the activity status of specified tenants. This tool updates the activity status, such as `ACTIVE` or `INACTIVE`, of one or more specified tenants within a collection (specified by `className`). Use this tool to manage tenant statuses efficiently within your system. Note: Understanding the request schema is necessary to properly create the stringified JSON input object for execution. This operation also requires path parameters. Modes: - GET_REQUEST_SCHEMA: Returns the schema. Only call if you don't already have it. Do NOT call repeatedly if you already received the schema. - EXECUTE: Performs the operation with the provided request body JSON. Note: You must also provide the required path parameters when executing. If you need the schema, call with mode='get_request_schema' ONCE, then execute.
Validate a data object's structure against the schema. Use this tool to check if a data object conforms to the specified collection schema and metadata rules without storing it. A successful validation returns a confirmation, while errors provide detailed feedback. Note: Understanding the request schema is necessary to properly create the stringified JSON input object for execution. Modes: - GET_REQUEST_SCHEMA: Returns the schema. Only call if you don't already have it. Do NOT call repeatedly if you already received the schema. - EXECUTE: Performs the operation with the provided request body JSON. If you need the schema, call with mode='get_request_schema' ONCE, then execute.