> SpecStream is the live web data engine for teams and agents. This is the Markdown copy of https://specstream.ai/docs/integrations/dive. Every page, with what SpecStream does and costs: https://specstream.ai/llms.txt

# Dive

Add SpecStream to Dive as a Streamable HTTP server in its MCP settings JSON. Any model you run in Dive can then search the live web, read pages and build datasets of structured rows, with a source URL and read time on every value.

[SpecStream for DiveAI assistants](https://specstream.ai/integrations/dive.md)

## What you need

-   Dive, the desktop MCP host from OpenAgentPlatform, with a model that supports function calling.
-   A SpecStream account. New accounts get $10 of credit on sign-up, no card.

1.  1

    ## 1. Create an API key

    In your SpecStream workspace, open [Agents & API](https://specstream.ai/workspace/agents?tab=keys&utm_source=dive), choose **API keys**, then **Create API key**, and give it a spending limit. Dive sends the key as a header on every call, so the server starts with no sign-in window.

2.  2

    ## 2. Open the MCP settings

    In Dive, open **Tools Management(MCP)** from the sidebar and select **Add / Edit MCP Config**. You can paste JSON in the editor (**MCP Tools Settings**), or select **\+ Add Config** and fill the same fields: name, transport, URL and a header under **Add Headers**.

3.  3

    ## 3. Add SpecStream to mcpServers

    Add a `specstream` entry beside the servers you already have. `transport` must be `streamable`: Dive's host accepts `stdio`, `sse`, `websocket` and `streamable`, and SpecStream serves Streamable HTTP. Put your key in place of `YOUR_SPECSTREAM_API_KEY`, then select **Save**.

    Dive MCP settings

    ```json
    {
      "mcpServers": {
        "specstream": {
          "transport": "streamable",
          "url": "https://api.specstream.ai/mcp",
          "headers": {
            "Authorization": "Bearer YOUR_SPECSTREAM_API_KEY"
          },
          "enabled": true
        }
      }
    }
    ```

4.  4

    ## 4. Check that it connected

    Back in **Tools Management(MCP)**, the SpecStream entry lists its tools with a switch for each one. Turn off any tool you do not want the model to call. If the entry says **Start failed, error occurred**, check that the key follows `Bearer` and that the URL ends in `/mcp`.

5.  5

    ## 5. Ask in a chat

    Pick a model with function calling. The **Tools** button in the chat shows how many tools are on, SpecStream's among them. Then ask:

    In a Dive chat

    ```text
    Search the web with SpecStream: which open-source vector databases released a new version this month? Give the version and the release page for each.
    ```

## What the model can call

Dive lists every tool the server offers. These are the ones a chat uses most:

-   `search_web` answers a question from the live web, with the results it read and a source URL and read time on each value.
-   `read_pages` returns the main text of up to 20 web pages or documents (PDF, Word, Excel) in one call.
-   `compile` reads a request for a dataset into conditions and columns, and `quote_capacity_work` prices it. Neither spends anything.
-   `authorize_capacity_work` starts the run, only with the maximum you approved. `get_find` and `list_find_rows` follow it and read the rows.

## Next to Dive's built-in Fetch

Dive ships a Fetch tool that reads an address the model already has. SpecStream finds the pages first with `search_web`, then reads up to 20 at once with `read_pages`. A page that refuses a plain read is tried again as a browser would ask for it, then in a real browser, then from the site's own feed, then from its newest public archive capture.

## Sign in with OAuth instead of a key

Leave out `headers` and select **Save**. On a Streamable HTTP entry that needs authorization, Dive shows **Connect**: select it, sign in to SpecStream in the browser, and approve the access. Requests then run on the account that signed in, with its balance and spending limits.

Dive's README marks its MCP server authentication as unstable and says it may ask you to authorize again. The key in a header keeps working until you revoke it.

## Cost

Connecting SpecStream to Dive is free. Each `search_web` call is one Search API call at $5 per 1,000 searches of up to 10 results, and each page `read_pages` returns is $0.80 per 1,000 pages; a page with no text is not charged. Both come out of your balance with no quote.

A dataset costs 5¢, 15¢ or $1 a row, depending on how much each row needs, plus 5¢ for a person's email and 15¢ for a phone. It runs only after you approve the quoted maximum, and a run that finds nothing costs nothing. [Balance, quotes, and billing](https://specstream.ai/docs/billing.md) has the details.

## Disconnect

In **Tools Management(MCP)**, open the menu on the SpecStream entry and select **Delete**, or turn it off. To stop the key everywhere, revoke it under **Agents & API**. Results you saved stay in your workspace. The [MCP guide](https://specstream.ai/docs/mcp.md) covers every tool. Questions go to [support@specstream.ai](mailto:support@specstream.ai).

## More

[SpecStream for Dive](https://specstream.ai/integrations/dive.md) covers what it does in Dive on one page. Every other tool has its guide in [Connect your tools](https://specstream.ai/docs/integrations.md). You see the most a request can cost before it starts, and the matches are paid from your balance. [Billing](https://specstream.ai/docs/billing.md) has the details. For help, email [support@specstream.ai](mailto:support@specstream.ai).