Connectors

to the systems you use.

Each connector is a node. Drop it into an agent like any other step, with the same guardrails and audit trail.

Ready-made connectors

Connectors for the tools you already use

Each group is a set of nodes in the agent editor. The same nodes work in every deployment.

Databases and warehouses
Query the systems that hold your records as a step in a workflow, and pass the result to the next node. Your data stays where it is.
Search, knowledge and governance systems
Use the search and knowledge platforms your organization already runs as a source, with the same grounding checks as any other.
Cloud storage and content management
Read from and write to the storage and document libraries that hold your policies, contracts and procedures, including saving a generated document back.
Development and collaboration tooling
Issue trackers, code hosting, boards and team messaging, so answers and updates land where your team already works.
Workplace productivity suites
Mail, calendars, documents and files, so an agent can prepare and file work in the tools people already have open.
Email and file transfer
Send mail through your own SMTP server and move files over SFTP on a schedule.
Web search and retrieval
Search and fetch public web sources as a step you add on purpose. Results are screened on the way in like any other outside content.
Audit destinations
Send the audit stream to object storage, Splunk, Datadog or your own collector, so the record sits with the rest of your security data.

The node library

Each card is a node in the agent editor. Drop it on the canvas, fill in the query and connection, and the next node can read the result.

MySQLConnects to a MySQL database, runs your SQL query and returns the rows as plain sentences or as CSV for the next node to use.

For everything else

When there is no ready-made node for a system, these let your team connect it quickly.

The REST node
Anything with an HTTP API is reachable as a step: URL, headers, body and operation set on the node, credentials pulled from the secrets you keep on the Configure tab.
Agents over MCP
Your agents are available as MCP tools, so any MCP client your team runs can call them, with the same guardrails and run record.
Models as cards
A model is set up the same way as a connector: a card, with the functions it serves ticked. Swap the model and everything built on top keeps working.

Start from a template

Open a template, point it at your data and change what you need. Every node is visible and editable.

What security teams ask before AI agents reach their systems of record.

Where does our data go when we connect a system?
A connector runs where NeuralSeek runs. In your own environment, that means inside your network. Host your own model too, and every call stays in-house.
Where do the credentials live?
On the Configure tab, as named secrets. A workflow refers to a secret by name and its value stays hidden. You insert it from a menu in the editor.
Can an agent write to a system as well as read?
Yes, through the same nodes. What each step sent and got back is in the run record, and a saved agent only runs when called by an API request or a schedule you set.