Agent access

Connect Faceless Radar to your agent

Skill, MCP, RSS and the REST API will read the same data: today's breakout faceless channels and videos, as our own fields plus YouTube IDs and links. Anonymous and read-only, no account and no key.

Coming soon. A rankings API behind Radar is in development; the Skill, MCP server, feed and the open-source repo are not built. Per the YouTube API policy check, any feed carries our own fields (niche, Style, cluster, write-ups, first-seen date) plus video and channel IDs with youtube.com links. Titles, thumbnails and view counts are not in it: your agent resolves them with its own key. Nothing on this page is a live link until then.

Agent Skill

install once, never update again

Preview of the final shape · not live yet

  1. Copy this and send it to your agent

    It installs the skill itself and will not touch other skills you have installed.

    Install the FacelessVideo Radar skill: https://facelessvideo.co/radar-skill/README.md
    After installing, request https://facelessvideo.co/api/radar/rankings?kind=channels&format=short&tier=1m once from the environment you run commands in; if it cannot connect, tell me how to allow network access.
    Tell me whether I should open a new session.
  2. Open a new session

    Most agents read skills at the start of a session, so the current conversation may not see it yet.

  3. Ask it something
    Which faceless Shorts channels created in the last month broke out this week, and what style do they use?

MCP

Add one server to your MCP client; it exposes the channel and video rankings, the breakdowns and the clusters as tools.

{
  "mcpServers": {
    "faceless-radar": {
      "url": "https://facelessvideo.co/api/mcp"
    }
  }
}

RSS

One feed per day of breakouts, with a link to each channel breakdown.

https://facelessvideo.co/radar/feed.xml

REST API

anonymous GET · our own fields and IDs
GET https://facelessvideo.co/api/radar/rankings?kind=channels&format=short&tier=1m&sort=nth&limit=20
Response shape
{
  "source": "FacelessVideo Radar",
  "retrievedAt": "2026-11-03T06:00:00Z",
  "expiresAt": "2026-11-04T06:00:00Z",
  "kind": "channels",
  "format": "short",
  "tier": "1m",
  "sort": "nth",
  "channels": [
    {
      "id": "UC…",
      "url": "https://www.youtube.com/channel/UC…",
      "niche": "history",
      "style": [
        "slideshow"
      ],
      "cluster": {
        "slug": "quiet-history",
        "name": "Quiet history slideshows"
      },
      "oneLiner": "…",
      "firstSeenAt": "2026-11-01",
      "topVideos": [
        {
          "id": "…",
          "url": "https://www.youtube.com/watch?v=…"
        }
      ]
    }
  ]
}

kind = channels | videos · format = short | long · tier = 1m | 2m | 3m (channels) · window = 7 | 30 (videos) · sort = nth | channelViews | subscribers | newest · technique, cluster, category, limit ≤ 50.

Resources

coming soon · not live
  • llms.txtSite description for language modelsSoon
  • Agent guideWhat the Skill reads; any agent can read it tooSoon
  • OpenAPI 3.1Full definition of the REST APISoon
  • Skill packageInstall notes and all filesSoon
  • GitHub · open-source datasetOur breakdowns and clusters, without YouTube dataSoon