An exhaustive Claude Code skill for the Olostep web scraping and data extraction API. Once installed, Claude Code can act as a principal engineer when integrating Olostep — it knows every endpoint schema, async orchestration flow, pagination paradigm, and edge case without you pasting docs.
Works with Claude Code and any agent that supports skills.
Contributions welcome! Open a PR or an issue.
| Endpoint | What Claude knows |
|---|---|
| Scrapes | Full request/response schema, actions, parsers, LLM extraction, screen sizes |
| Answers | AI agent with structured JSON output, NOT_FOUND handling, sources |
| Batches | Step-by-step async orchestration: submit → poll → list items → retrieve |
| Crawls | Step-by-step async orchestration: kickoff → poll → list pages → retrieve |
| Maps | Domain URL discovery with glob filtering and cursor pagination |
| Searches | Semantic web search with deduplication |
| Retrieve | Shared content extraction endpoint for batches/crawls |
| Files | Context upload lifecycle |
| Schedules | Serverless cron for any endpoint |
| Webhooks | Completion notifications with retry behavior |
| Metadata | Key-value pairs with PATCH semantics |
| SDKs | Python (sync + async) and Node.js patterns |
Personal (available across all your projects):
git clone https://github.com/olostep-api/olostep-claude-code-skill ~/.claude/skills/olostepProject-scoped (shared with anyone using the repo):
git clone https://github.com/olostep-api/olostep-claude-code-skill .claude/skills/olostepClaude Code auto-detects the skill — no restart needed.
Other Methods
# Git submodule
git submodule add https://github.com/olostep-api/olostep-claude-code-skill .claude/skills/olostepUpdating
# If cloned directly
cd ~/.claude/skills/olostep && git pull
# If using git submodule
git submodule update --remote .claude/skills/olostepOnce installed, Claude Code automatically loads the skill when you're working with Olostep. You can also invoke it directly:
/olostep
Or just ask naturally — the skill triggers when Claude detects Olostep-related work:
"Scrape this URL and extract the product title and price as JSON"
"Set up a batch job to scrape these 500 Google Search URLs with the parser"
"Crawl docs.example.com and retrieve markdown for all pages under /guides/"
PRs and issues are welcome. If you notice an endpoint change, a missing field, or want to add examples, please open a pull request.