Skip to main content

Product tour

LightNow is the MCP control plane for AI clients. It brings server discovery, trust signals, Runtime Profiles, client connection and runtime visibility into one operating model. Your AI client keeps one LightNow Proxy entry while the selected profile defines the MCP servers behind it.

MCP server workflowThe path from discovery to client setup
LightNow Registry server browser with filters, MCP server cards and trust scores.
1BrowseFind MCP candidates

Filter by source, category, transport, trust and capability signals before a server reaches an agent.

LightNow Redis MCP Server detail page with metadata, version tabs and README content.
2InspectEvaluate one version

Read upstream README context beside version, repository, capability and trust evidence.

LightNow Redis MCP Server integrations tab with install command and client setup sections.
3ConnectTranslate into setup

Use integrations for experiments; move repeat use into Runtime Profiles and Local Proxy.

The operating loop

Teams normally repeat four actions as MCP becomes part of daily development:

  1. Discover a server and inspect its source, transport and capabilities.
  2. Decide whether its ownership and trust evidence fit the intended use.
  3. Add it to a personal or organization Runtime Profile.
  4. Connect an AI client and verify the resulting runtime.
Operating loopProduct surfaces map to the same loop your team runs when MCP becomes daily infrastructure
1RegistryDiscover candidates

Browse uses source, category, transport, trust, favorites and capability filters before a server is considered.

Server catalog
2DetailEvaluate one version

Summary, capabilities, trust, integrations and versions keep the upstream story beside LightNow evidence.

Integration details
3ProfileCentralize repeat use

Runtime Profiles store the server set that client sync, export and Local Proxy resolve later.

Runtime Profiles
4OperateVerify local posture

Config status, proxy health and runtime events separate client config drift from upstream failures.

Config status + health

Discover and assess servers

The Registry helps you compare MCP servers without starting from scattered READMEs and copied config snippets. Browse by source, category, transport, capabilities and trust signals, then open one version for the details that matter to your decision.

Capabilities and trust answer different questions:

  • Capabilities describe what the server can expose to an AI client.
  • Trust signals describe what LightNow knows about the publisher, source, endpoint and evaluated version.

Treat those signals as decision support, not as a substitute for your own risk assessment.

Build a reusable Runtime Profile

A Runtime Profile is the named server set for a workflow. Use a personal profile for your own clients or an organization profile when a team should share and govern the same set.

Before connecting a client, make sure every profile server has the command, URL, headers, environment values or secret references it needs. The profile can exist while still being incomplete; readiness must be resolved before rollout.

Connect an AI client

For supported clients, the recommended setup writes one LightNow entry that starts LightNow Proxy in the same execution environment as the client. The proxy resolves the selected Runtime Profile and routes MCP requests to its stdio and Streamable HTTP upstreams.

This keeps per-server commands and secret values out of the normal client file. Direct sync and bridge exports remain compatibility paths for workflows that need a native client configuration.

Govern and observe the runtime

Organization settings can select a default profile, choose managed clients and decide whether unmanaged client entries may remain. Apply those settings with policy-driven sync only after the profile works for a pilot client.

LightNow Proxy can report health and metadata-only runtime events. These events help distinguish client-config drift, profile resolution failures and upstream errors without storing tool arguments, tool results, resource contents, workspace paths, secrets or authorization headers.

Choose your next step