Crawlora for OpenClaw

Give your OpenClaw agent 23 tools for live structured web data: search, product research, maps, transcripts, finance, SEC filings, and company hiring.

This is a native OpenClaw tool plugin, backed by the official @crawlora-org/sdk. Scraping runs on Crawlora's hosted service. You need a Crawlora account and API key; API requests consume your account's credits and rate limits. The plugin itself is MIT-licensed.

Install

Requires OpenClaw 2026.9.3 and Node >=24.16.0 <25 or >=26.1.0. This initial release declares the exact OpenClaw version tested; broader compatibility will be added after testing.

openclaw plugins install clawhub:@crawlora-org/openclaw-plugin --accept-capabilities

The capability flag approves registration of the 23 tools listed below. Review the tool list and data-handling section before installing.

Add the following to your existing OpenClaw configuration, merging the plugins fields:

{
  "plugins": {
    "entries": {
      "crawlora": {
        "enabled": true,
        "config": { "apiKey": "${CRAWLORA_API_KEY}" }
      }
    }
  }
}

Set CRAWLORA_API_KEY in the environment of the process running your OpenClaw Gateway, then restart it. Alternatively, set apiKey in the plugin config directly. Plugin config takes precedence over the environment; the field is marked sensitive in OpenClaw's configuration UI.

If you use a plugin allowlist, add crawlora to the existing plugins.allow array. Keep any other entries. If you restrict agent tools, also allow this plugin's tools using your existing tool policy.

Verify registration:

openclaw plugins inspect crawlora --runtime --json

Try asking:

• “Search Google for recent battery recycling research.”
• “Find coffee shops in San Francisco using Crawlora.”
• “Get Apple's annual SEC financials and recent filings.”
• “Search Amazon for noise cancelling headphones.”
• “Show Stripe's public Greenhouse jobs.”

Tools

All tool names start with crawlora_ to avoid collisions with core tools or MCP integrations. This is a curated adapter, not the full Crawlora endpoint catalog. For the full catalog, connect the hosted Crawlora MCP server.

Parameter details

• Google search takes q and optional num (10–100); uses US/English results.
• Google News, Google Videos, and Bing take q and optional count (1–50).
• SEC company tools require ticker or cik. crawlora_sec_financials accepts statement: income, balance, cash_flow; period: annual, quarterly; and limit: 1–20. Filings allow 1–500, insider transactions 1–30.
• Hiring signals supports the curated providers greenhouse (token), lever (company), ashby (org), smartrecruiters (company), and workday (tenant, datacenter, site). More providers exist in Crawlora's full API.
• Greenhouse token means a public board slug, not an authentication token.
• The YouTube tool requests the API's default JSON transcript. Availability depends on the video and upstream captions.
• Google Trends uses the API defaults for geography and time window.

Behavior and data handling

Tools return { "data": ... }, preserving the SDK response inside data. This avoids interpreting domain-specific status fields as OpenClaw execution status. Scraped content is external data and must not be treated as agent instructions.

Requests go only to https://api.crawlora.net/api/v1; CRAWLORA_BASE_URL is deliberately ignored. Each call uses the current API key, a 60-second timeout, and the active tool's cancellation signal. Automatic retries are disabled to avoid repeating credit-consuming calls. HTTP failures become short, actionable tool errors; raw upstream error bodies, headers, and API keys are not included.

The plugin has no shell execution, local file access, background jobs, or telemetry. Tool parameters and the API key are transmitted to Crawlora to fulfill requests. The API key is supplied in the x-api-key header, never a tool parameter.

Development

npm ci
npm run plugin:validate
npm test
npm pack --pack-destination /tmp
openclaw plugins install npm-pack:/tmp/crawlora-org-openclaw-plugin-1.0.0.tgz --force --accept-capabilities
openclaw plugins inspect crawlora --runtime --json

plugin:build compiles TypeScript, generates openclaw.plugin.json, and adds catalog/UI metadata. The package ships built JavaScript and the manifest. Tests use the real OpenClaw registration helper and published Crawlora SDK with mocked HTTP to verify every tool's URL, method, body, and authentication, plus cancellation, configuration changes, and failure handling.

Optional live smoke test, using your own account credits:

CRAWLORA_API_KEY=... npm run test:live

Release

From a clean committed checkout:

npm ci
npm run plugin:validate
npm test
clawhub package validate . --openclaw-version 2026.9.3
clawhub package publish . --owner crawlora-org --dry-run
clawhub package publish . --owner crawlora-org --wait

ClawHub's security checks must finish before a release is publicly installable. Publish credentials stay outside this repository.

License

MIT. Copyright 2026 Crawlora.