# Wire Discovery

Give Wire a topic and it searches the web for related content on a schedule, pulling the results into a container as entries your agents can query. It's a standing feed: set up "OpenAI model releases" or "Federal Reserve rate decisions" once, pick how often it refreshes, and the container stays current without you doing anything. Wire runs the search infrastructure, so there are no API keys or accounts to set up.

## How it's organized

Wire Discovery splits in two, the same shape as every other connector:

1. **The connector** is enabled once at the **organization** level. There's nothing to configure and no credentials to enter, so it's a single click that turns the feature on for your org.
2. **A topic** is a slice of the web that flows into a specific container. You add topics on a container's **Sources** tab. One org can feed many containers, each with its own topics.

## Permissions

| Action | Who can do it |
|---|---|
| Enable / disable Discovery at the org level | **Owner**, **Admin** |
| View the connectors list | All members |
| Add a topic to a container | Anyone with **edit** access to that container |
| Edit, refresh, or remove a topic | Anyone with **edit** access to that container |

## Step 1. Enable Discovery for your org

1. Open **Organization → Connectors** in the Wire dashboard.
2. Click **Add connector** and pick **Wire Discovery**.
3. Click **Enable Discovery**. That's it, there are no keys to enter.

## Step 2. Add a topic to a container

1. Open the container you want to feed.
2. Go to the **Sources** tab.
3. In the **Wire Discovery** row, click **Setup** (or **Edit** if you already have topics).
4. Click **Add topic** and fill in the basics:
   - **Topic**: enter it the way you'd type it into a search box, a concrete subject Wire can search for, e.g. `NVIDIA earnings` or `EU AI regulation`.
   - **Refresh**: how often Wire checks for new content (see below).
5. Optionally open **Advanced options** for finer control (also below).
6. Click **Add topic**. Wire runs the first discovery immediately.

Each topic shows its cadence, its preferred sites (if any), and when it last refreshed, along with a status pill:

| Pill | Meaning |
|---|---|
| **Pending** | Queued, the first run hasn't started yet |
| **Syncing** | A run is in progress, content is landing |
| **Synced** | Caught up |
| **Error** | Something went wrong, hover the row for details |

### Refresh cadences

| Option | Behavior |
|---|---|
| **Once** | Pull once now and never re-check. You can still refresh it by hand later. |
| **Every 30 minutes** | For fast-moving topics where minutes matter. |
| **Hourly** | |
| **Every 6 hours** | |
| **Daily** (default) | Checks once a day. |

After the first run, every refresh only pulls what's **new since the last check**, so a 30-minute topic asks for "what's appeared in the last half hour" rather than re-pulling everything.

### Advanced options

| Option | What it does |
|---|---|
| **Preferred sites** | Restrict results to specific sites, e.g. `reuters.com, techcrunch.com`. Up to 20. Leave empty to search everywhere. |
| **Region** | Bias results toward a country or city. |
| **Content type** | Narrow to news, research, company pages, or financial content. Defaults to anything. |
| **First run covers** | How far back the initial fill reaches (past day / week / month / any time). Later runs ignore this and use the since-last-check window. |
| **Results per check** | The most results to pull each run. |

## Editing and refreshing a topic

Open the **Sources** tab, hit **Edit**, then the pencil on a topic to change its query, cadence, or any advanced option in place. Saving updates the topic; the next run uses the new settings. To pull right now, hit the refresh icon on the topic row.

## What lands in the container

Each discovered item becomes an entry tagged `discovered`, carrying the article title, a link back to the source, and the body when Wire can read it. From there it's processed like any other content: chunked, embedded, and woven into the container's knowledge graph, so your agents reach it through the standard tools (`wire_search`, `wire_navigate`, and friends).

Wire never duplicates the same article: if the same URL turns up again, or two of your topics surface it, it stays a single entry.

## Removing a topic and disabling Discovery

Hit **Edit** on the Sources tab, then the trash icon on a topic to stop discovering for it. **Content already discovered stays in the container** until you delete it; removing a topic only stops future runs.

Disabling Wire Discovery at the org level (**Organization → Connectors → Disable**) stops every topic across all your containers from refreshing. Again, content already discovered stays put. You can re-enable later, but topics need to be re-added.

## Costs

Discovery uses credits as it runs. You're charged per refresh, plus an amount for each new item a run brings in, so a topic that refreshes more often, or pulls more results, uses more credits. The analysis of discovered content (chunking, embeddings, entity extraction) is free, like every other write. See the [pricing page](https://usewire.io/pricing/) for the current per-refresh and per-item rates.

## Reading paywalled or protected sites

Wire reads publicly available content. It isn't designed to access paywalled or login-gated pages, and it doesn't try to defeat bot protection. For sites like those, a topic will still capture the headline, the link, and whatever summary is publicly visible, but not the full article body. For the same reason, it may skip social media sites, where posts usually sit behind logins and aggressive bot protection. This is deliberate: Discovery is a good-citizen reader, not a scraper that fights walls.

## Troubleshooting

**A topic shows Synced but I expected more results**
Discovery only returns what's genuinely new since the last run, and only what's publicly readable. A quiet topic, a narrow set of preferred sites, or paywalled sources will all reduce what lands. Widen the topic, drop the site restriction, or lower the cadence's expectations.

**The "Add topic" form says Discovery isn't enabled**
Discovery is enabled per organization. Ask an owner or admin to enable it under **Organization → Connectors**, then add topics from the container.

**An article I care about didn't get the full body**
The source is most likely paywalled or bot-walled. Wire captures the headline, link, and any public summary, but won't pull the gated body. That's expected.