Agent skills

scrapeio

by xhoantran

npx skills add xhoantran/scrapeio-agent --skill scrapeio -a claude-code -y

SKILL.md

ScrapeIO returns public social media data as the same JSON shape on every platform. Every call spends credits; cached and failed calls are free.

Setup

The key is in the SCRAPEIO_API_KEY environment variable. If it is missing, ask the user to create a free key at https://scrapeio.com/login and set it. Never print the key.

Two ways to call

MCP tools (preferred when the scrapeio MCP server is connected): call the platform tool directly, for example tiktok_profile with { "handle": "nike" }. For anything without its own tool, use list_endpoints to find a path, then call_endpoint.
HTTP: fetch https://scrapeio.com/llms.txt for the endpoint list (or https://scrapeio.com/llms-full.txt for every parameter), then:

curl -s "https://scrapeio.com/v1/tiktok/profile?handle=nike" -H "x-api-key: $SCRAPEIO_API_KEY"

Rules

Choose the narrowest endpoint for the question. Prefer handle when the user gives a username.
The answer is in data. Lists have data.items, data.cursor and data.has_more.
Fetch more pages only when the user needs them; each page is a separate paid call.
credits_used and credits_remaining come back on every response. Tell the user when credits run low.
403 means private and 404 means not found: report it, do not retry. 502 means every source is down: retry once after a few seconds. 402 means the account needs credits (https://scrapeio.com/pricing).

Report

Lead with the numbers the user asked for. Name posts by caption or title and date, and link to them with the url field. Field names match across platforms, so one table can compare TikTok, Instagram and YouTube side by side.

View raw Markdown

---
name: scrapeio
description: Pull public social media data with the ScrapeIO API - profiles, posts, videos, comments, transcripts, search, trending content and ad libraries from TikTok, Instagram, YouTube, X (Twitter), Facebook, Reddit, LinkedIn and 25 more platforms. Use when the user asks for creator or account stats, post metrics, comments, transcripts, hashtag or keyword research, trends, influencer discovery, or competitor ads.
---

# ScrapeIO

ScrapeIO returns public social media data as the same JSON shape on every platform. Every call spends credits; cached and failed calls are free.

## Setup

The key is in the `SCRAPEIO_API_KEY` environment variable. If it is missing, ask the user to create a free key at https://scrapeio.com/login and set it. Never print the key.

## Two ways to call

1. **MCP tools** (preferred when the `scrapeio` MCP server is connected): call the platform tool directly, for example `tiktok_profile` with `{ "handle": "nike" }`. For anything without its own tool, use `list_endpoints` to find a path, then `call_endpoint`.
2. **HTTP**: fetch https://scrapeio.com/llms.txt for the endpoint list (or https://scrapeio.com/llms-full.txt for every parameter), then:

```bash
curl -s "https://scrapeio.com/v1/tiktok/profile?handle=nike" -H "x-api-key: $SCRAPEIO_API_KEY"
```

## Rules

- Choose the narrowest endpoint for the question. Prefer `handle` when the user gives a username.
- The answer is in `data`. Lists have `data.items`, `data.cursor` and `data.has_more`.
- Fetch more pages only when the user needs them; each page is a separate paid call.
- `credits_used` and `credits_remaining` come back on every response. Tell the user when credits run low.
- 403 means private and 404 means not found: report it, do not retry. 502 means every source is down: retry once after a few seconds. 402 means the account needs credits (https://scrapeio.com/pricing).

## Report

Lead with the numbers the user asked for. Name posts by caption or title and date, and link to them with the `url` field. Field names match across platforms, so one table can compare TikTok, Instagram and YouTube side by side.