Skip to content

Explore the public reference platform

platform.mcpruntime.org is MCP Runtime's public reference deployment. With a user account and team access, you can publish and deploy a sample MCP server, create grants, connect a client, and explore the user and team views. You do not need your own Kubernetes cluster. Allow about 10 minutes.

This platform is for getting a feel for the user and team workflows. To try platform setup and administration, install MCP Runtime in your own cloud or on-premises environment; see Platform Installation.

One limit: the public platform cannot verify which agent makes the example tool calls, so it denies them. You can still try the deployment, grant, connection, and dashboard workflows below.

1. Install the CLI

curl -fsSL https://raw.githubusercontent.com/mcp-runtime/mcp-runtime/main/install.sh | MCP_RUNTIME_OS=darwin sh

The installer detects your CPU and installs the binary to ~/.local/bin.

curl -fsSL https://raw.githubusercontent.com/mcp-runtime/mcp-runtime/main/install.sh | MCP_RUNTIME_OS=linux sh

The installer detects your CPU and installs the binary to ~/.local/bin.

irm https://raw.githubusercontent.com/mcp-runtime/mcp-runtime/main/install.ps1 | iex

The installer uses a user-local directory and adds it to your user PATH. Open a new terminal after installation.

Pin a specific release by setting MCP_RUNTIME_VERSION to its tag before running the installer.

The macOS/Linux installer logs detection, download, and installation steps, shows download progress, and retries failed transfers. Color is automatic in a terminal; set NO_COLOR=1 to disable it.

Browse all binaries on the latest GitHub release. If macOS or Linux cannot find mcp-runtime, add ~/.local/bin to your shell's PATH, for example with export PATH="$HOME/.local/bin:$PATH".

Verify:

mcp-runtime --version

To upgrade the CLI, run the installer again and verify mcp-runtime --version. The CLI binary and the public reference platform are separate release tracks: the CLI changes only when a new release is published; the UI at platform.mcpruntime.org is updated by its operator and does not need a local install.

2. Log in

Use an account on the public reference platform. You need team access to deploy the sample server. Ask for a user account and team membership if you do not have them; this walkthrough does not require a platform administrator account.

mcp-runtime auth login \
  --api-url https://platform.mcpruntime.org \
  --email you@example.com --password '...' \
  --profile me

mcp-runtime auth status    # confirm the profile is active

3. Deploy an example server

This example needs Git, Go 1.26+, and Docker with its daemon running. These are example build prerequisites; installing the release CLI does not require Go.

Clone the repo to get the example server source:

git clone https://github.com/mcp-runtime/mcp-runtime
cd mcp-runtime/examples/oauth-example-go-2025-11-25

Run it locally to discover its tool names, then scaffold the metadata:

go run . &
SERVER_PID=$!
# Wait for the listening log and confirm http://localhost:8088/health.

mcp-runtime server init workspace-demo \
  --from-server http://localhost:8088
# Discovered: aaa-ping, add, create_task, draft_release_note, echo, lower, slugify, upper, whoami

kill $SERVER_PID

Validate the metadata, build the image, push it, and deploy:

mcp-runtime server validate --metadata-dir .mcp

mcp-runtime server build image workspace-demo --tag v1
# Prints the exact image ref, e.g.: registry.mcpruntime.org/myteam/workspace-demo:v1

mcp-runtime server push \
  --image registry.mcpruntime.org/myteam/workspace-demo:v1 \
  --scope tenant

mcp-runtime server deploy workspace-demo --scope tenant --metadata-dir .mcp

server push publishes the local image through the authenticated platform API.

Confirm it is running:

mcp-runtime server list

4. Grant access and connect

Create a grant that allows an agent to call echo and add. Replace myteam with your actual team slug. First choose an active agent ID from your team's directory:

mcp-runtime agent list myteam --status active
# If your team has no suitable agent, ask a team owner to create one.
AGENT_ID=agt_01arz3ndektsv4rrffq69g5fav # replace with the ID from the list

Agent display names such as Cursor are not IDs. Use the immutable ID returned by mcp-runtime agent create or shown by mcp-runtime agent list. Members see only agents covered by an applicable grant or their own active session; ask a team owner for an ID if the list is empty.

mcp-runtime access grant init workspace-cursor \
  --server workspace-demo \
  --namespace mcp-team-myteam \
  --agent-id "$AGENT_ID" \
  --tool echo \
  --tool add \
  --output grant.yaml

mcp-runtime server validate --metadata-dir .mcp --grant-file grant.yaml
mcp-runtime access grant apply --file grant.yaml

With the grant applied, start the adapter proxy to connect an MCP client to your server. The grant must exist before the platform can issue the agent's session:

mcp-runtime adapter proxy \
  --runtime-url https://mcp.mcpruntime.org/workspace-demo/mcp \
  --server workspace-demo \
  --agent "$AGENT_ID" \
  --auto-refresh \
  --listen 127.0.0.1:8099

Point Claude Desktop, Cursor, or any MCP client at http://127.0.0.1:8099. The client can connect on the public platform. If it calls echo or add, the platform denies the call because it cannot verify which agent made it. The error is 401 missing_identity. See Client Connections for the identity setup needed to allow these calls on your own installation.

5. See it in the analytics

Open platform.mcpruntime.org and go to Analytics → Tools. You can inspect recorded calls and their decisions. Calls blocked because the platform could not verify the agent appear as denied.

In Server Catalog or My Activity, confirm the deployed server is visible to your account. Open its details to review the endpoint and connect configuration. The UI uses the same deployment and policy state as the CLI. The Gateway badge identifies servers configured to run the Runtime gateway; readiness and OAuth/header authentication have their own badges.

Platform admins can open Usage analytics to filter recorded gateway requests by time range, public/org/team scope, team, server, MCP method, tool, policy decision/version and HTTP outcome. Choose Teams, search a team by name or slug, then select it to scope the chart and totals. The Server requests table combines method and tool activity, with request counts, HTTP errors, p95 latency and a link to the latest request trace per group. Its rows are the top selected number of groups; summary totals and the chart cover all matching recorded requests. The chart shows local time and requests per second. HTTP failures, latency, identified users/agents, sessions and response traffic come from recorded gateway events; servers without a gateway are not counted.

The console sidebar groups pages under Runtime, Workspace, and Platform, according to your account's access. Use Filter navigation to find a page and the account control at the bottom to view your account or sign out. On compact screens, open the navigation menu in the header. The theme button switches between dark and light mode.

What's next