← Registry

Data Collection

stormy.ai

This MCP server provides tools for searching social media profiles, retrieving video transcripts, finding contact emails, and estimating pricing for data lookup services.

1 endpoint11 known toolsFirst detected August 8, 2026Last detected August 8, 2026

ENDPOINT 1

https://stormy.ai/mcp

No auth detected

MCP server metadata

Name
Stormy Social Data
Version
1.29.0
Capabilities
experimentalpromptsresourcestools
Server instructions

Use Stormy to search and look up public social data across Instagram, YouTube, TikTok, X, LinkedIn, and Reddit. - A people search reads and judges real profiles, so it takes roughly 45-90 seconds. Tell the user that before you call it, and do not treat a wait as a hung tool. Past 100 seconds the call returns a job_id instead of results; poll it. - Use a durable job for any people search you expect to be large or slow, and a direct call for a single known profile. A direct search is fine for an ordinary request; it is the same engine either way. - When the user names a size — "under 50k followers", "people my size" — pass min_followers/max_followers rather than relying on the wording alone, and say the words in `query` as well. - A short result is an answer. `warnings` says whether the search was thin because the niche is thin or because part of the machine did not answer; read it before you retry. - Estimate first when the user asks about cost or requests 25 or more paid results. - Give each new job a stable idempotency_key derived from the user's intent. If a call is retried, reuse that key; never start a replacement for a nonterminal job. - Poll get_social_job only after poll_after_seconds. queued, scheduled, running, throttled, and retrying are nonterminal; stop only when terminal=true. - Search and profile tools never return contact data. Call find_emails only when the user explicitly asks for emails or contacts; only successful verified finds are billed. Email targets are limited to 25 per call or durable job, so split larger lists into batches of at most 25 with a distinct stable idempotency_key for each batch. - Supported platforms are Instagram, YouTube, TikTok, X, LinkedIn, and Reddit. For any other network, do not attempt a search; return the supported choices and ask the user to pick one. - Never request or place API keys in tool arguments. On upgrade_required or insufficient_balance, return the supplied upgrade_url. On rate_limit, wait for retry_after_seconds instead of retrying immediately. Return source URLs and the exact usage receipt when supplied by a tool. Treat unavailable platform fields as null, not as facts to infer.

Known tools 11

search_people

Search one social network and return normalized public results, source URLs, and a usage receipt.

Inferred read-only
get_video_transcript

Everything said in a YouTube video, as text.

Inferred read-only
lookup_profile

Look up one profile URL, handle, or channel ID and return normalized public fields plus a usage receipt.

Inferred read-only
find_emails

Find verified contact emails for 1-25 Instagram, TikTok, or YouTube profiles.

Inferred read-only
estimate_price

Return the authoritative rate card and maximum workflow estimate without spending credits.

Inferred read-only
account_status

Show entitlements, remaining prepaid usage, and the checkout or top-up URL.

Inferred read-only
describe_social_data

Return the machine-readable platform, field, pricing, reliability, and workflow contract.

Inferred read-only
start_social_job

Queue fresh, bulk, or email work with durable retries and status.

Potential side effects
get_social_job

Return status, progress, exact poll timing, result/error, and the complete timeline.

Inferred read-only
list_social_jobs

List recent durable jobs, optionally filtered by any documented job status.

Inferred read-only
cancel_social_job

Cancel a queued, scheduled, throttled, or retrying job before a worker starts it.

Inferred read-only

CONNECT WITH APPROVAL

Client installation

Review this server and its permissions before adding it. Secret placeholders must be set locally.

Codex

~/.codex/config.toml

[mcp_servers.stormy-social-data]
url = "https://stormy.ai/mcp"
enabled = true
Claude Code

.mcp.json

{
  "mcpServers": {
    "stormy-social-data": {
      "type": "http",
      "url": "https://stormy.ai/mcp"
    }
  }
}
Claude Desktop

Settings → Connectors → Add custom connector

Name: stormy-social-data
Remote MCP URL: https://stormy.ai/mcp

Add this remote URL as a custom connector in Claude Desktop. Availability depends on the user plan and workspace policy.

Cursor

.cursor/mcp.json

{
  "mcpServers": {
    "stormy-social-data": {
      "url": "https://stormy.ai/mcp"
    }
  }
}
Visual Studio Code

.vscode/mcp.json

Add to Visual Studio Code
{
  "servers": {
    "stormy-social-data": {
      "type": "http",
      "url": "https://stormy.ai/mcp"
    }
  }
}
Generic MCP

Client-specific MCP configuration

{
  "name": "stormy-social-data",
  "transport": "streamable-http",
  "url": "https://stormy.ai/mcp"
}
MCP Inspector

Run the official MCP Inspector locally and enter the indexed Streamable HTTP endpoint.

TRUST AND VERIFICATION EVIDENCE

Trust Data Available

BuiltWith Trust API v2 evidence for stormy.ai was fetched 2026-08-09T22:06:40.099Z.

Trust status Trusted

stormy.ai is assessed as Trusted: Domain has an established technology history spanning over a year.

Indexed

Evidence is source-attributed and does not guarantee that a third-party server is safe. Risk labels are conservative metadata heuristics.