Question 1

Do I need to know how to code?

Accepted Answer

No. Paste a URL, describe in plain English what you want to track, and Trawl emails you when it changes. The code editor and API are there for developers; optional for everyone else.

Question 2

What kind of data can Trawl watch?

Accepted Answer

Public pages your account is permitted to access: prices, availability, stock, job listings, competitor promos, property listings, tender notices. Trawl encourages users to respect target site ToS and robots.txt, and ships warnings when something looks off. Built-in API rate limits cap how often any account can trigger scraps, which indirectly limits target site load; per-host throttling depends on your schedule, and the responsibility for what you scrape stays with you. See our [Acceptable Use Policy](/aup).

Question 3

How does AI auto-fix actually work?

Accepted Answer

When a run fails, Trawl re-reads the current page, matches the new layout against a **shared knowledge base of cross-tenant patterns**, and ships a patch to your selectors. Successful patterns feed back into the library, so the next site that breaks the same way is fixed even faster.

Question 4

Can I drive Trawl from Claude Code or an MCP agent?

Accepted Answer

Yes, on the **Free plan** — no credit card needed. `npm install -g @trawlme/cli && trawl login` logs you in, then `trawl token` prints your session token. Pass it to Claude Desktop, Cursor, or any MCP client as a Bearer token: `Authorization: Bearer <token>`. The MCP server is available to all users. **API keys** (for stable, long-lived agent connections that never need re-auth) are available on **paid plans** (Growth and Pro) — the session token is re-authenticated every 7 days.

Question 5

What does the AI Digest email look like?

Accepted Answer

A morning bullet list of what actually changed overnight: price deltas, stock transitions, new listings. Plus a 7-day sparkline per scrap. No noise, no unchanged rows, no marketing.

Question 6

How does smart proxy routing work?

Accepted Answer

Trawl handles all the routing intelligence so you don't have to. Each run tries the least intrusive (and cheapest) legitimate path first, and only escalates if the page fails to render. Week after week, runs that succeed are quietly downgraded to cheaper tiers. Your cost trends down as long as the site stays cooperative. Shared knowledge across the platform helps speed up that learning curve. Included on every plan.

Question 7

Can I consume Trawl data from my own infra?

Accepted Answer

Yes. Every scrap exposes a **live JSON endpoint** you can consume from your own workers. Use webhooks to push results straight into your pipeline, or the CLI to pull data on-demand into any CI job.

Question 8

Do you store my scraped data?

Accepted Answer

Only the last extraction result, to help debug. Nothing more. Data flows to your webhook or API consumer: **your control**, **your retention policy**.

Question 9

Is monitoring public web pages legal?

Accepted Answer

Monitoring publicly available pages is generally legal in most jurisdictions. Trawl respects `robots.txt` and rate limits by default, and our [Acceptable Use Policy](/aup) sets out what users are responsible for (target site ToS, data protection laws, no circumvention of technical protection measures). For regulated use cases, consult your legal team. **We ship the tool, you own the policy.**

Question 10

Is Trawl free to start?

Accepted Answer

Yes. The Free plan gives you a one-shot 500-compute signup grant (no card, no renewal) — enough to try monitors, smart proxy routing, anti-bot resolution, the AI suite, and the **MCP server and CLI**. Connect Claude Desktop, Cursor, or any MCP-aware agent on day one.

API reference