Skip to content

On IBM Cloud

on IBM Cloud

Buy NeuralSeek from the IBM Cloud catalog. Run it as SaaS or on your own OpenShift, with watsonx models and watsonx Assistant.

Start nowAsk about your setup

Buy onIBM Cloud

How to buy

Buy from the catalog or run your own

The product and its professional services are both in the IBM Cloud catalog. You can also run the container on your own OpenShift.

Every deployment has the same features, however you buy it.

  • The product in the IBM Cloud catalog

    Provision NeuralSeek from the IBM Cloud account you already run, under the terms the catalog entry links.

    Open the catalog entry

  • Professional Services in the same catalog

    Onboarding, training and implementation, bought as blocks of hours through the same catalog and procurement route as the software.

    Open the services entry

  • A container on your own OpenShift

    The same software as containers on OpenShift or Kubernetes, in the account, network and region you choose, under a container license. Talk to us to set it up.

    Ask about deployment

Native IBM integrations

The IBM services NeuralSeek uses

Built-in integrations. Add a governed application layer to the watsonx, Discovery and Assistant setup you already run.

NeuralSeek

  • watsonx.ai as the model layer

    A supported model provider. Pick a watsonx model at setup, or a different one for each step of an agent, and change it later in configuration.

  • Watson Discovery and watsonx Discovery as knowledge bases

    Recommended for relevance tuning and dynamic filter queries, which boost the right documents and limit a query to what each user may see. watsonx Discovery adds vector and hybrid search.

  • watsonx Assistant as the virtual agent

    A quick-start template adds fallback search: when the assistant has no answer for a question, NeuralSeek answers from your knowledge base. Curated answers export into the assistant and are flagged when sources change.

  • watsonx.governance from inside an agent

    An agent step logs a retrieval and its answer to watsonx.governance, or sends it a raw payload, so IBM's model record and NeuralSeek's audit trail describe the same request.

Inside an agent

Governance inside every run

Each integration is a step inside a run: a model is chosen per step, a knowledge base answers, and the governance call happens while the answer is produced.

console › governance › runs
A run in the console with each step and the controls applied to it.
A governed run in the console, with each step and its controls.

watsonx.ai in the model layer

Use watsonx.ai models your way

Add a watsonx.ai model as a card in the configuration, choose the functions it serves, and add a second provider beside it if you like. When prices or models change, you update the configuration and your agents keep running.

console › neural config › llm details
The LLM Details panel in the console: an Add an LLM button, a note on load balancing, and two model cards side by side, each with its connection info and a grid of platform functions ticked per model.
Each model is a card, with the jobs it handles checked.

On IBM today

Built on the IBM stack customers already run

NeuralSeek working beside watsonx.ai in production.

  • Retail & commercial banking

    A customer-facing digital banking assistant with guided loan applications.

    • A customer assistant for general digital banking questions
    • Loan applications with careful handling of customer data
    • Internal system integration governed by NeuralSeek Guardrails
    • Every regulated customer interaction under the same policy
    Digital banking · Loan applications

Story 1 of 1: NatWest

Where it runs

Your account, your region, your controls

Run NeuralSeek from the catalog or on your own OpenShift, in the region you choose, alongside watsonx.

NeuralSeek on AWS and Azure runs the same container with the same features.

  • One container, the same features

    Provisioned from the catalog or self-hosted on your own OpenShift, it's the same software with the same controls.

    Deployment options

  • A region you choose

    Stored data can be kept in a region, and the container route puts the whole deployment in the account and region you pick.

    Data and AI sovereignty

  • With watsonx

    watsonx supplies models, data and model-level governance. NeuralSeek adds the governed application layer on top.

    NeuralSeek and IBM watsonx

Security review answers

Short answers for your security review.

  1. Is data encrypted?

    Yes, in transit and at rest. Everything that flows to or through the service is encrypted over TLS, and everything stored is encrypted too.

  2. Is our data used to train anything?

    No. Neither NeuralSeek nor any of its subprocessors use customer data to train models or systems, and generated answers belong to the customer.

  3. What happens to personal data in a request?

    Content flagged as personal data is handled by the action you configure: flag, mask, hide or delete. Requests that carry personal data for a personalized answer are handled the same way.

Audited and compliant

On the container route NeuralSeek runs inside the OpenShift or IBM Cloud account you already approved, under your network controls, key management, logging and incident process.

Audited

SOC 2 Type II

Audited every year.

NeuralSeek is SOC 1 and SOC 2 Type II audited and HIPAA compliant.

See certifications for details.

Next step

Docs, support and terms

The documentation, support and terms the IBM Cloud catalog entries link to.

Deploying on your own OpenShift? Talk to us and we'll help you set it up.

  • Documentation

    Integration references and setup guides for watsonx.ai, Discovery, watsonx Assistant and watsonx.governance.

    Read the documentation

  • Terms of service

    The full terms linked from the catalog entry.

    Read the terms

Start building with NeuralSeek today.

Sign up online and start with a two-week trial, or talk to a NeuralSeek expert about running it in your own cloud.

Start nowRequest a demo