Demo-ready suite cockpit
Integration Matrix
AgentUI is the human-facing interface layer. These services remain integrations with their own identities; unavailable runtime snapshots are labeled degraded.
Full demo flow
PlatPhorm AgentUI Suite Demo: Root -> MCP -> Spec -> Sandbox -> Evals -> AgentUI
AgentUI creates preview_ready packets by default. Protected target sends require PLATPHORM_API_KEY and completion is shown only when the target service returns a result reference.
Discover trusted topology from PlatPhorm root
Start with the network graph so AgentUI renders trusted-suite context instead of hardcoded site assumptions.
- Endpoint
- https://platphormnews.com/api/network/graph
- Handoff
- root to root · ready
- Auth
- public-safe read/preview
- Result ref
- not confirmed yet
Select MCP tool schema rendering after topology is visible.
Render MCP tool form in AgentUI
MCP remains the capability registry; AgentUI renders tool input schemas and JSON-RPC previews.
- Endpoint
- https://mcp.platphormnews.com/api/mcp
- Handoff
- agentui to mcp · ready
- Auth
- public-safe read/preview
- Result ref
- not confirmed yet
Validate the rendered contract with Spec before any target execution.
Validate interface contract with Spec
Spec owns contract truth. AgentUI prepares the schema packet and displays returned validation only when Spec confirms it.
- Endpoint
- https://spec.platphormnews.com/api/v1/validate
- Handoff
- agentui to spec · ready
- Auth
- public-safe read/preview
- Result ref
- not confirmed yet
Prepare a Sandbox command packet as a dry-run handoff.
Prepare Sandbox dry-run command packet
Sandbox owns bounded command execution. AgentUI prepares the command packet and keeps it preview_ready unless Sandbox confirms execution.
- Endpoint
- https://sandbox.platphormnews.com/api/sandbox/receive-handoff
- Handoff
- agentui to sandbox · preview_ready
- Auth
- PLATPHORM_API_KEY for send
- Result ref
- not confirmed yet
Send the UI and workflow packet to Evals as a scorecard preview.
Prepare Evals release-gate packet
Evals owns scoring. AgentUI shows scorecards only after Evals confirms a result.
- Endpoint
- https://evals.platphormnews.com/api/evals/send-handoff
- Handoff
- agentui to evals · preview_ready
- Auth
- PLATPHORM_API_KEY for send
- Result ref
- not confirmed yet
Confirm AgentUI can receive proof handoffs from the suite.
Confirm AgentUI receive-handoff readiness
AgentUI owns the human-operable proof receiver and does not claim persistence or completion without a result reference.
- Endpoint
- https://agentui.platphormnews.com/api/agentui/receive-handoff
- Handoff
- agentui to agentui · ready
- Auth
- public-safe read/preview
- Result ref
- not confirmed yet
Run live read-only probes if the demo network is available.
MCP Registry
Canonical MCP capability registry and federated tools/resources/prompts.
https://mcp.platphormnews.com- OpenAPI
- configured
- MCP
- configured
- llms-index
- configured
Source: environment. Operator actions: configured server-side.
- MCP remains the canonical capability registry; AgentUI renders MCP tool schemas.
platphormctl mcp validate agentui --best-effort
- OpenAPI
- configured
- MCP
- degraded
- llms-index
- configured
Source: environment. Operator actions: configured server-side.
No fresh runtime snapshot is stored in AgentUI yet.
platphormctl site inspect agentui --best-effort
- OpenAPI
- configured
- MCP
- degraded
- llms-index
- configured
Source: environment. Operator actions: configured server-side.
platphormctl harness run spec-evals-browserops-loop --dry-run
Webhooks
Webhook payload contracts, simulation plans, and delivery verification.
https://webhooks.platphormnews.com- OpenAPI
- configured
- MCP
- degraded
- llms-index
- configured
Source: environment. Operator actions: configured server-side.
QSTASH runtime signals are not configured; AgentUI can still render webhook payload forms without claiming delivery.
platphormctl site inspect agentui --best-effort
- OpenAPI
- configured
- MCP
- degraded
- llms-index
- configured
Source: environment. Operator actions: configured server-side.
- Evals scorecards are displayed only when Evals returns a scorecard.
platphormctl harness run spec-evals-browserops-loop --dry-run
- OpenAPI
- configured
- MCP
- degraded
- llms-index
- configured
Source: environment. Operator actions: configured server-side.
- Core DB/QStash signals enable packet visibility only; Sandbox execution is not claimed without a target result.
platphormctl harness run spec-evals-browserops-loop --dry-run
BrowserOps
Browser journey, accessibility, and screenshot validation.
https://browserops.platphormnews.com- OpenAPI
- configured
- MCP
- degraded
- llms-index
- configured
Source: environment. Operator actions: configured server-side.
- Browserbase signals enable browser proof handoff readiness; BrowserOps validation is not claimed until confirmed.
platphormctl browserops validate-tool-ui mcp get_health --dry-run
- OpenAPI
- configured
- MCP
- degraded
- llms-index
- configured
Source: environment. Operator actions: configured server-side.
No fresh runtime snapshot is stored in AgentUI yet.
platphormctl site inspect agentui --best-effort
- OpenAPI
- configured
- MCP
- degraded
- llms-index
- configured
Source: environment. Operator actions: configured server-side.
QSTASH runtime signals are not configured; AgentUI can still render generated status panels without claiming monitor checks.
platphormctl site inspect agentui --best-effort
Trace
Trace/span propagation, evidence links, and Vercel metadata correlation.
https://trace.platphormnews.com- OpenAPI
- configured
- MCP
- degraded
- llms-index
- configured
Source: environment. Operator actions: configured server-side.
- Trace links are safe metadata; raw request fingerprints are not exposed.
platphormctl site inspect agentui --best-effort
- OpenAPI
- configured
- MCP
- degraded
- llms-index
- configured
Source: environment. Operator actions: configured server-side.
platphormctl site inspect agentui --best-effort
- OpenAPI
- configured
- MCP
- degraded
- llms-index
- configured
Source: environment. Operator actions: configured server-side.
No fresh runtime snapshot is stored in AgentUI yet.
platphormctl site inspect agentui --best-effort
- OpenAPI
- configured
- MCP
- degraded
- llms-index
- configured
Source: environment. Operator actions: configured server-side.
No fresh runtime snapshot is stored in AgentUI yet.
platphormctl site inspect agentui --best-effort
Fingerprint
Redaction, trust, and fingerprint summary review panels.
https://fingerprint.platphormnews.com- OpenAPI
- configured
- MCP
- degraded
- llms-index
- configured
Source: environment. Operator actions: configured server-side.
No fresh runtime snapshot is stored in AgentUI yet.
platphormctl site inspect agentui --best-effort
- OpenAPI
- configured
- MCP
- degraded
- llms-index
- configured
Source: environment. Operator actions: configured server-side.
No fresh runtime snapshot is stored in AgentUI yet.
platphormctl site inspect agentui --best-effort
- OpenAPI
- configured
- MCP
- degraded
- llms-index
- configured
Source: environment. Operator actions: configured server-side.
No fresh runtime snapshot is stored in AgentUI yet.
platphormctl site inspect agentui --best-effort
- OpenAPI
- configured
- MCP
- degraded
- llms-index
- configured
Source: environment. Operator actions: configured server-side.
No fresh runtime snapshot is stored in AgentUI yet.
platphormctl site inspect agentui --best-effort
- OpenAPI
- configured
- MCP
- degraded
- llms-index
- configured
Source: environment. Operator actions: configured server-side.
No fresh runtime snapshot is stored in AgentUI yet.
platphormctl site inspect agentui --best-effort
Markdown Tools
Markdown rendering and content transformation workflows.
https://markdown.platphormnews.com- OpenAPI
- configured
- MCP
- degraded
- llms-index
- configured
Source: environment. Operator actions: configured server-side.
No fresh runtime snapshot is stored in AgentUI yet.
platphormctl site inspect agentui --best-effort
- OpenAPI
- configured
- MCP
- degraded
- llms-index
- configured
Source: environment. Operator actions: configured server-side.
platphormctl site inspect agentui --best-effort
Layout
Responsive layout workbench preview, scoring, export, and design evidence.
https://layout.platphormnews.com- OpenAPI
- configured
- MCP
- degraded
- llms-index
- configured
Source: environment. Operator actions: configured server-side.
platphormctl agentui dry-run workflow.json --block layout_preview
ASCII
Image-to-ASCII conversion outputs that AgentUI can render, export, and publish as artifacts.
https://ascii.platphormnews.com- OpenAPI
- configured
- MCP
- degraded
- llms-index
- configured
Source: environment. Operator actions: configured server-side.
platphormctl agentui render-tool ascii convert --dry-run
PlatPhorm AI
Optional server-side Vercel AI Gateway or PlatPhorm AI configuration for model-assisted labels, summaries, search, and plans.
https://agentui.platphormnews.com- OpenAPI
- configured
- MCP
- degraded
- llms-index
- configured
Source: environment. Operator actions: configured server-side.
platphormctl agentui summarize-run run.json --model-assisted --dry-run
PlatPhorm CLI
platphormctl command documentation and validation harnesses.
https://cli.platphormnews.com- OpenAPI
- configured
- MCP
- degraded
- llms-index
- configured
Source: environment. Operator actions: configured server-side.
platphormctl agentui tools
PlatPhorm Root Graph
Network graph topology source for trusted domain discovery.
https://platphormnews.com/api/network/graph- OpenAPI
- degraded
- MCP
- degraded
- llms-index
- configured
Source: environment. Operator actions: configured server-side.
platphormctl site inspect agentui --best-effort