Connect compatible AI clients to public product knowledge or authenticated operational deployments. Select the server and domain, inspect sessions and question history, and use the tools available to your identity. Privileged actions follow the connected server’s permissions and operating policy.
Read product capabilities, formats and deployment guidance through typed MCP tools, resources and prompts. The public endpoint cannot access your private platforms or change their data.
Authorize an AI client against the operational deployment. Select the relevant domain and inspect the tools it exposes. Administrators can review activity, diagnostics and approval history where the server supports them.
Nexus includes a personal MCP setup screen and a guided operations interface for the resources available to the current account. These are separate from the public website’s product-information assistant.
Start with personal MCP setup
Review the connection identity, Streamable HTTP endpoint and credential-handling guidance before authorizing an AI client.
Choose a permitted resource
The guided resource picker covers campaigns, Search, XML, RTB, Video, CPA and publisher operations within the current role and module access.
About these screenshots
Captured from Nexus. Identifying details are hidden. The account name and endpoint are hidden. The operations capture shows resource selection only; no AI connection, command or data change was executed for these screenshots.
A new interface to AdTech
Product discovery and operations. Two clearly defined boundaries.
The assistant and MCP server use the same curated Nexus product model. The website helps buyers understand the platform; MCP lets AI clients discover that knowledge through typed tools and resources instead of scraping pages.
The operational workspace
One place to understand what your AI clients are doing.
Connect the MCP servers your business operates. The administration workspace exposes the capabilities each server supports, organized by deployment and domain—not mixed into one untraceable stream.
01
Multiple servers, one workspace
Register separate MCP deployments and select the server and domain before reviewing its operational context. Keep production and test environments distinct.
02
Questions, tools and diagnostics
Review session activity, question history, tool outcomes and latency. Inspect the evidence behind an interaction instead of relying on the final chat answer alone.
03
Governed actions and automation
Use the connected server’s approval, automation and audit capabilities. An AI request never replaces the server’s access checks or the deployment’s operating policy.
04
Learning with a review trail
Inspect learning proposals, supporting evidence and review outcomes where the remote server exposes them. Keep changes to operational guidance visible to administrators.
01get_nexus_overview
Discover the platform, AdTech engineering heritage, revenue engines and AI-native operating model.
02get_product_capabilities
Inspect typed capabilities for Search, feeds, OpenRTB, video or CPA.
03get_targeting_controls
Compare geographic, device, domain, sub-ID, filtering and traffic-limit controls across supported channels.
04get_ad_format_guide
Explain what an ad looks like, where it appears, how it is bought, targeted, delivered and measured.
05get_operating_capabilities
Research traffic quality, feed and bid policy, partner operations, direct campaigns, contextual monetization and custom engineering.
06recommend_deployment
Map a business model and channel mix to a phased Nexus deployment.
07get_security_model
Understand authentication, authorization and infrastructure boundaries.
Example questions
See how public product questions map to MCP tools.
AI CLIENT→NEXUS MCP→recommend_deployment
Which Nexus product fits an ad network?
Start with the channels already carrying revenue, map partner access and traffic policy, then phase Search, feeds, OpenRTB, video or CPA into one tenant-aware control plane.
Illustrative product answer · no customer data · no write operation
How AI clients connect
A governed path from question to platform context.
01Discover
The AI client connects to the Nexus Streamable HTTP endpoint and negotiates MCP capabilities.
02Inspect
It lists typed tools, reads the product catalog resource and can use the solution-discovery prompt.
03Reason
The client combines published Nexus product information with the customer’s business question.
04Act safely
Public tools are read-only. Connected operational servers expose privileged tools behind scoped authentication, deployment policy and approval requirements.
Open by design
One protocol. Multiple AI experiences.
OpenAI and Codex workflowsClaude-compatible MCP clientsCustom agents and copilotsIDE and desktop MCP clientsInternal automation systemsDirect Streamable HTTP clients
Compatibility depends on the client supporting the Nexus Streamable HTTP transport and its authentication model. Exact connection instructions are validated per client and deployment.
Permission model
Product discovery and authenticated operations.
PUBLIC
Product knowledge
Read-only tools, resources and prompts for capability and deployment discovery.
AUTHENTICATED
Tenant context
Identity-bound access to explicitly assigned platform and operational context.
APPROVED
Operational actions
Operational writes require server-side authorization; approval and audit requirements follow the connected server’s policy.
AI-ready from day one
Give your team—and its AI clients—a direct path to platform knowledge.