Qrawl
Qrawl
Qrawl is QUI's AI-native web browsing system. It lets characters search the web, extract page content, and focus on specific sections — all without browser automation. It's pure content extraction, not a headless browser.
Qrawl is patent pending.
What Qrawl Does
Qrawl provides five operations:
| Operation | Purpose |
|---|---|
| Search | Web search via Brave, Google CSE, or SearXNG |
| Skim | Quick page overview — headings, key links, structure |
| Read | Full page content extraction with readability processing |
| Focus | Targeted extraction of specific sections, elements, or content matching a query |
| Interact | Parse and submit HTML forms |
Results are cached so repeated access to the same page doesn't make redundant requests.
How Characters Use Qrawl
When the Qrawl node is enabled on a character, the character can trigger web operations:
[TRIGGER:qrawl:search:latest quantum computing breakthroughs]
[TRIGGER:qrawl:fetch:https://example.com/article]
The character receives the results and can discuss, summarize, or act on the content. In agentic mode, a character can chain multiple Qrawl calls — search for information, read the top results, then synthesize findings.
Qrawl vs Browser/Playwright
QUI has two separate web access systems:
| Feature | Qrawl | Playwright (via MCP) |
|---|---|---|
| Technology | Pure Python (httpx + BeautifulSoup) | Full browser automation |
| Speed | Fast — no browser startup | Slower — launches a browser instance |
| JavaScript | Not supported — static HTML only | Full JS rendering |
| Use case | Content extraction, research, reading articles | Interacting with web apps, testing, filling forms |
| Trigger | [TRIGGER:qrawl:...] |
[TRIGGER:mcp:playwright:...] |
Use Qrawl for reading and research. Use Playwright for interactive web applications that require JavaScript.
Enabling Qrawl
- Add the Qrawl node to your character's canvas in the Visual Builder
- The character gains access to web search and content extraction triggers
- In ThinkThing, use the dedicated Qrawl nodes (QRawl Search, QRawl Skim, QRawl Read, QRawl Focus)
Search Providers
Qrawl supports multiple search providers:
| Provider | Setup |
|---|---|
| Brave | Requires a Brave Search API key |
| Google CSE | Requires Google Custom Search Engine credentials |
| SearXNG | Self-hosted search aggregator (no API key needed) |
Configure the provider and credentials in the Qrawl settings or per-search via parameters.
Rate Limiting
Web searches are rate-limited per character to prevent abuse. Default: 30 searches per day. Configurable per character and per time period (1 hour, 6 hours, day, week, month).
Content fetching (skim, read, focus) is not rate-limited but results are cached for 30 minutes.
