Scrape TikTok, YouTube, and Instagram. Comprehend video with multimodal AI. Deliver structured understanding to your agent via MCP or REST — in one call.
From raw data to structured comprehension — the full pipeline, managed and metered.
TikTok, YouTube, and Instagram — search, profiles, hashtags, videos, comments, and posts. 15+ structured scrape actions with rate-managed Apify actors.
Gemini and DashScope multimodal models understand video content — scenes, transcripts, entities, emotional arcs, and brand-safety flags with confidence scores.
One-line install. Your agent gets scrape, comprehend, and billing tools via the Model Context Protocol. Streamable HTTP — no SDK, no glue code.
Pay as you go. Credits are minor-unit integers — transparent, metered per action class. Set spend caps, buy packs, configure auto-topup.
Stream comprehension results as they're generated via Server-Sent Events. Partial results with depth control — raw or pre-digested insights.
Per-tool metrics, cost tracking, and margin monitoring. Know which actors are healthy, what each call costs, and where to optimize.
No pipelines to build. No models to host. Just point your agent at a URL.
Call a scrape action — search, profile, hashtag, video, comments. Apify actors fetch structured data from TikTok, YouTube, or Instagram.
Pass any media URL to understand_media. Gemini/DashScope multimodal models produce a structured, confidence-scored comprehension.
Your agent receives the result as a tool call response — or via REST. One-line MCP install, streaming supported, credits deducted transparently.
Paste any public TikTok, YouTube, or Instagram URL. We'll fetch, comprehend, and structure it. Rate-limited per IP.
Pay for what you use. Credits never expire. No infrastructure fees.
500 credits / month
5,000 credits / month
Volume credits + SLA
No SDK. No glue code. Your agent gets the tools it needs.
# Add to your MCP client config (Claude, Cursor, etc.)
npx @social-connectors/mcp --url https://matane.pawbytes.io/mcp --key sc_live_yourkey# Your agent now has these tools:
tools:
- scrape_tiktok_search # { queries: ["skincare"], resultsPerPage: 10 }
- understand_media # { url: "https://tiktok.com/@x/video/123" }
- account # { } → check balance
# Example: comprehend a TikTok video
{
"tool": "understand_media",
"input": {
"url": "https://www.tiktok.com/@glossy/video/7280...",
"depth": "raw"
}
}
# Returns structured JSON:
{
"media_ref": { "platform": "tiktok", "identity_type": "content_hash" },
"scenes": [{ "start_time": 0, "description": "..." }],
"transcript": { "full_text": "..." },
"entities": { "products": [{ "name": "..." }] },
"brand_safety_flags": { "is_safe": true },
"confidence": 0.92
}