FourA Integration

Use the FourA integration to give assigned Dawn agents web scraping and exploration tools for public pages, structured responses, and HTTP APIs.

What FourA is for

FourA gives Dawn agents a controlled web request layer for pages and HTTP APIs that are useful during research, troubleshooting, support, competitive analysis, and day-to-day operations.

Use it when the agent needs to fetch a public page, read its visible text, extract links or structured fields, call an HTTP API endpoint, or submit a login-free form — for tasks that are not already covered by another Dawn integration.

FourA is different from Dawn web search. Web search is for finding relevant pages across the web. FourA is for inspecting a specific URL or HTTP endpoint once the agent knows what it should look at.

Before you start

  • A Dawn workspace where you are an owner or manager.
  • A FourA API key.
  • The FourA API base URL, if your account uses a non-default host.
  • A clear idea of which agents should be allowed to use web scraping and exploration tools.

Keep the FourA API key in the connection. Do not paste it into prompts, memory pages, or channel messages.

Connect FourA in Dawn

  1. Open Integrations in Dawn.
  2. Click Add connection.
  3. Choose FourA.
  4. Enter the FourA API key.
  5. Leave the default API base URL unless your FourA account requires a different one.
  6. Validate the connection.

After validation, Dawn creates or shows a single source called Web scraping and exploration.

FourA does not require you to link individual websites or scrape projects.

The integration exposes one source: Web scraping and exploration. Link that source, then assign it to the agent profiles that should use it.

Connecting FourA does not automatically expose the tools to every agent. Agents only receive the FourA tools after the source is assigned to their profile.

What agents can do

Assigned agents can inspect specific URLs and HTTP APIs as part of normal conversations.

Typical requests include:

Open this product page, summarize the visible copy, and list the main calls to action.
Check this public changelog URL and tell me what changed since the last release notes draft.
Inspect this API endpoint, parse the JSON response, and put the main fields in a table.
Read the pricing page for these three competitors and compare the plan names, entry prices, and stated limits.

The agent can use normal web request controls — methods, headers, request bodies, redirect handling, response validation, JSON parsing — and proxy routing when the task requires it.

Two request modes

Each FourA request runs in one of two modes:

  • Direct request — a fast HTTP fetch with browser-like fingerprinting (via curl-impersonate), so the request looks the way a normal browser request would.
  • Proxy routing — the request is routed through FourA’s proxy network with anti-bot handling for cases where a direct path would be blocked, rate-limited, or unreliable.

Most workflows should start with the direct mode. Use proxy routing only when the page or endpoint requires it — it carries different access expectations and consumes more provider capacity. Dawn does not run a headless browser, so JavaScript-rendered pages where the content only appears after scripts run are out of scope for this integration.

Safety and boundaries

FourA is a webpage scraping capability, not a way to bypass access controls.

Do not use it for paywalled content, private pages, systems that require someone else’s credentials, or sites where automated access would violate the site’s terms. Dawn also applies platform protections around target URLs, response size, timeout handling, and secret redaction.

FourA results are used in the conversation where the agent calls the tool. Dawn does not turn every scraped page into a long-lived knowledge source automatically.

Troubleshooting

  • Connection fails: confirm the FourA API key is valid and the base URL is correct.
  • Tools are missing: confirm the Web scraping and exploration source is linked and assigned to the active agent profile.
  • A URL fails: check whether the target URL is public, reachable, and suitable for automated access.
  • A response is too large: ask for a narrower page section, a specific JSON field, or a smaller result set.
  • Proxy behavior is unexpected: retry without proxy routing first, then use proxy routing only if the direct request path is unsuitable.
  • A page renders empty: the integration does not support headless browsers yet, so JavaScript-rendered content that only appears after scripts run will not be in the response. Use a different source for those pages.
Validation Prompts

Use these prompts in Dawn chat to quickly validate this setup area. 

Check whether this agent has FourA web exploration access and run a safe known-page read test.
Use FourA to inspect this public URL and summarize the title, links, and visible text.
Completion Checklist
  • FourA API key validated.
  • Web scraping and exploration source linked.
  • Source assigned to the intended agent profile.
  • Agent can fetch and summarize a known public page.
  • Proxy routing reviewed before relying on it for a workflow.