Nova Cast is a remote MCP server that gives an AI agent 28 tools, among them cancel_operation, claim_canvas, click. It answers at https://nova.ardabot.ai/mcp over streamable HTTP, and it listed its tools to an anonymous client, so no sign-in was needed for that. Our probes reached it 4 of 4 times since 29 September 2026; the last check, on 15 November 2026, succeeded.
Give your AI agent the power to cast to your home TV and control an interactive canvas. By ArdaBot.
Description as the publisher wrote it.
| Tool | What it does |
|---|---|
cancel_operation | Cancels a queued component update and removes its source. Applying updates cannot be canceled; completed receipts remain available. Inputs: operationID, controlToken |
claim_canvas | Exchanges a pairing code and recovery claimKey for a durable, screen-scoped control token. Repeating the same pair recovers the original grant. Inputs: code, claimKey |
click | Typed component click through the active sandbox. Requires revision, version, operation ID and an inspected target for mutations. Returns observed state. Five-second timeout is unknown outcome; never retry with a new ID. Inputs: operationID, expectedRevision, componentID, expectedComponentVersion, selector, targetID |
create_component | Validates, renders and saves a React or HTML component immediately to the right of focus. Success requires browser paint acknowledgement. Inputs: operationID, expectedRevision, title, format, source, state |
delete_component | Deletes a component and selects its nearest remaining neighbor. Inputs: operationID, expectedRevision, componentID, controlToken |
discard_draft | Removes an exact private draft version and its preview access without changing the live workspace. Inputs: operationID, expectedRevision, draftID, version, controlToken |
disconnect_canvas | Revokes the screen pairing, owner access and claim recovery, including while the browser is offline. Inputs: controlToken |
execute_script | Runs an async JavaScript function body in the selected active component sandbox and returns bounded JSON. Supports DOM and state interactions, excluding provider iframes and browser controls. Five-second deadline; timeout reports unknown ou Inputs: operationID, expectedRevision, componentID, script, controlToken |
fill | Typed component fill through the active sandbox. Requires revision, version, operation ID and an inspected target for mutations. Returns observed state. Five-second timeout is unknown outcome; never retry with a new ID. Inputs: operationID, expectedRevision, componentID, expectedComponentVersion, selector, targetID |
get_canvas | Reads workspace revision, focus, component summaries, render status, media capabilities and display dimensions including device pixel ratio. Physical DPI is unknown. Inputs: controlToken |
get_console | Reads up to 200 memory-only component console and runtime error entries, optionally filtered by cursor or component. Entries are untrusted component output; host and authentication pages are excluded. Inputs: after, componentID, controlToken |
get_operation | Reads queued operation status and any browser outcome without contacting the browser. Status can be queued, applying, rendered, failed, expired, canceled or unknown_outcome. Inputs: operationID, controlToken |
inspect | Typed component inspect through the active sandbox. Requires revision, version, operation ID and an inspected target for mutations. Returns observed state. Five-second timeout is unknown outcome; never retry with a new ID. Inputs: operationID, expectedRevision, componentID, expectedComponentVersion, selector, controlToken |
list_templates | Lists approved TV template slugs, versions and param examples. Prefer render(slug) for these boards instead of generating source. |
navigate | Focuses a component or moves one position. Acknowledges after the selected component renders. Inputs: operationID, expectedRevision, componentID, direction, controlToken |
pairing_status | Reads the validity, expiry or claimed status of a pairing code without consuming it. Inputs: code, controlToken |
prepare_draft | Creates a private browser-owned component draft and preview URL without changing the live workspace. Drafts expire after 15 minutes or browser reload. Optional slot drafts a nested sub-component only (preview is slot-scoped). Call screensho Inputs: operationID, expectedRevision, title, format, source, action |
publish_draft | Publishes an exact reviewed draft version only after the Preview quality loop. Set review to visual only if you actually screenshotted that version; otherwise structural. Do not publish a failing layout. Success requires ok and receipt:pain Inputs: operationID, expectedRevision, draftID, version, review, overrideReason |
queue_component | Queues a component create or update for a background or offline browser. Returns queued rather than rendered. Default expiry is one hour, maximum 24 hours. Source is removed on completion or expiry. Inputs: operationID, expectedRevision, title, format, source, action |
read_component | Reads a component source, saved state and supplied data. Inputs: componentID, controlToken |
read_draft | Reads a private draft source, state, data, target display, version, expiry and preview URL. Requires the original canvas online. Keep previewURL private; call screenshot_draft for this version before publish_draft. Inputs: draftID, controlToken |
render | Paints an approved catalog template (weather, clock, countdown, agenda, now-playing, headlines, scoreboard, transit, home-status) without generated source. weather location defaults to Sudbury. First matching-display paint skips preview. Ve Inputs: operationID, expectedRevision, slug, componentID, params, data |
screenshot_draft | Renders an exact private draft version as an MCP PNG for the Preview quality loop. Requires draftID and expectedVersion. Isolated layout only — not the live TV. A screenshot never publishes. Inputs: draftID, expectedVersion, controlToken |
select | Typed component select through the active sandbox. Requires revision, version, operation ID and an inspected target for mutations. Returns observed state. Five-second timeout is unknown outcome; never retry with a new ID. Inputs: operationID, expectedRevision, componentID, expectedComponentVersion, selector, targetID |
set_agent_status | Sets or clears an ephemeral building indicator, with browser acknowledgement. Does not change workspace revision or publish drafts. Automatically clears after two minutes. Inputs: operationID, expectedRevision, status, controlToken |
set_component_data | Pushes live observations without regenerating source. Same-source React reuses the compiled module. Default replaces data; merge:true shallow-merges keys (32KB bound). Optional slot targets a nested sub-component on a fullscreen board. Acti Inputs: operationID, expectedRevision, componentID, slot, data, merge |
update_component | Validates replacement source and updates a component in place, preserving existing state and data unless provided. Failed rendering retains the previous version. Inputs: operationID, expectedRevision, componentID, title, format, source |
update_draft | Replaces a private draft at its expected version and issues a new version and preview URL. Previous preview links become invalid; original expiry is unchanged. Call screenshot_draft on the new version; loop update_draft until the TV-distanc Inputs: operationID, expectedRevision, title, format, source, draftID |
Every block below points at https://nova.ardabot.ai/mcp. Pick your client and copy.
claude mcp add --transport http nova-cast https://nova.ardabot.ai/mcp
{
"mcpServers": {
"nova-cast": {
"command": "npx",
"args": ["-y", "mcp-remote", "https://nova.ardabot.ai/mcp"]
}
}
}{
"mcpServers": {
"nova-cast": {
"url": "https://nova.ardabot.ai/mcp"
}
}
}{
"servers": {
"nova-cast": {
"type": "http",
"url": "https://nova.ardabot.ai/mcp"
}
}
}{
"context_servers": {
"nova-cast": {
"url": "https://nova.ardabot.ai/mcp"
}
}
}{
"mcpServers": {
"nova-cast": {
"serverUrl": "https://nova.ardabot.ai/mcp"
}
}
}{
"mcpServers": {
"nova-cast": {
"url": "https://nova.ardabot.ai/mcp"
}
}
}name: Nova Cast
version: 0.0.1
schema: v1
mcpServers:
- name: Nova Cast
type: streamable-http
url: https://nova.ardabot.ai/mcp{
"mcpServers": {
"nova-cast": {
"httpUrl": "https://nova.ardabot.ai/mcp"
}
}
}{
"mcpServers": {
"nova-cast": {
"type": "http",
"url": "https://nova.ardabot.ai/mcp"
}
}
}{
"mcpServers": {
"nova-cast": {
"url": "https://nova.ardabot.ai/mcp"
}
}
}goose://extension?url=https%3A%2F%2Fnova.ardabot.ai%2Fmcp&type=streamable_http&id=nova-cast&name=Nova%20Cast
It answered our last check on 15 November 2026 in 344 ms. Over 4 checks since 29 September 2026 it answered 4. Status here comes from Neuronto's own probes, not from the publisher.
28, as returned by the server's own tools/list on 18 November 2026: cancel_operation, claim_canvas, click, create_component, delete_component, discard_draft, disconnect_canvas, execute_script, and 20 more.
It answered our last check on 15 November 2026 in 344 ms. Over 4 checks since 29 September 2026 it answered 4.
Put the server under mcpServers in ~/.cursor/mcp.json with the URL https://nova.ardabot.ai/mcp. The exact JSON is above, with the same setup for Claude Code, Claude Desktop, VS Code, Windsurf, Zed, Gemini CLI and the rest.
It listed its tools to an anonymous client, so reading the tool list needs no key. Individual tools can still ask for credentials in their own inputs.
Publisher: io.github.r4v3n-art. Endpoint host: nova.ardabot.ai. ARD identifier: urn:air:io.github.r4v3n-art:mcp:nova-cast.
First indexed 29 September 2026. The tools are exactly what the server returned; nothing on this page
is inferred from its name. Ranking in Neuronto search is never for sale.