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

# eve

Add SpecStream to an eve agent as an MCP connection. The agent searches the live web, reads pages and asks for structured lists through SpecStream's MCP server, and the key stays in your environment, out of the conversation.

[SpecStream for eve, Agent frameworks & builders](https://specstream.ai/integrations/eve.md)

## What you need

-   An eve project (`npx eve@latest init my-agent`) on Node.js 24 or later.
-   A SpecStream account. New accounts start with $10 of credit, no card. [Create one](https://specstream.ai/signup?utm_source=eve).
-   A SpecStream API key with a spending limit, which step 1 creates.

1.  1

    ## 1. Create an API key

    In your SpecStream workspace, open **Agents & API**, choose **API keys**, then **Create API key**. Turn on **Authorize paid work**: a key without it can't search or read pages. Set a **Total spending limit** above $0, because that limit is what caps the spending of everything that uses the key.

    Copy the key when it's shown. It isn't shown again.

2.  2

    ## 2. Add the key to the environment

    Add the key to the Vercel project that hosts the agent, then pull it into your local environment. Locally, a line `SPECSTREAM_API_KEY=...` in `.env.local` works too.

    Terminal

    ```bash
    vercel env add SPECSTREAM_API_KEY
    vercel env pull
    ```

3.  3

    ## 3. Add the connection

    Create one file under `agent/connections/`. The file name is the connection name, so this one registers as `specstream`. `auth.getToken` sends the key as `Authorization: Bearer`, and eve calls it on each connection, so the key is read from the environment and never enters the model's context. `tools.allow` keeps the agent to search and page reading:

    agent/connections/specstream.ts

    ```typescript
    import { defineMcpClientConnection } from "eve/connections";
    
    export default defineMcpClientConnection({
      url: "https://api.specstream.ai/mcp",
      description:
        "SpecStream: search the live web, read web pages and documents, and return each answer with the page it came from.",
      auth: {
        getToken: async () => ({ token: process.env.SPECSTREAM_API_KEY! }),
      },
      tools: { allow: ["search_web", "read_pages"] },
    });
    ```

4.  4

    ## 4. Ask

    Run `npx eve dev` and ask about something recent. The model finds the tools with `eve__search` and calls them as `specstream__search_web` and `specstream__read_pages`. Check that the turn shows a SpecStream tool call.

    In the terminal UI

    ```text
    Use SpecStream to find what changed in the latest Node.js release, read the release notes, and name the page you used.
    ```

## What each tool does

The SpecStream server gives the agent three kinds of work. Add a tool's name to `tools.allow` to give it to the agent.

| Tool | What it does | Cost |
| --- | --- | --- |
| `search_web` | Searches the live web with `{"q": "<the question>"}` and returns up to 10 results, each with its page. | $5 per 1,000 searches |
| `read_pages` | Returns the main text of 1 to 20 pages or documents (PDF, Word, Excel), read live, or from SpecStream's own dated copy when a page is slow or refuses, with `{"urls": [...]}`. | $0.80 per 1,000 pages |
| `compile`, `quote_capacity_work`, `authorize_capacity_work` | Deep Search: reads a request in plain words, prices it, and starts it after a maximum is approved. Rows come back with `list_find_rows`. | 5¢, 15¢ or $1 a delivered row, plus 5¢ for a person's email and 15¢ for a phone |

## Deep Search with a person's approval

To let the agent ask for structured lists, allow `compile`, `quote_capacity_work`, `authorize_capacity_work`, `get_find` and `list_find_rows`. `authorize_capacity_work` starts paid work with the approved maximum, so make the turn wait for a person before it runs:

agent/connections/specstream.ts

```typescript
tools: {
    allow: [
      "search_web",
      "read_pages",
      "compile",
      "quote_capacity_work",
      "authorize_capacity_work",
      "get_find",
      "list_find_rows",
    ],
  },
  approval: ({ toolName }) =>
    toolName === "specstream__authorize_capacity_work" ? "user-approval" : "not-applicable",
```

## Deploy

Run `eve deploy`, or push to the Git-connected Vercel project. The deployed agent reads `SPECSTREAM_API_KEY` from the project's environment, so keep the variable in every environment that runs the agent.

## Cost

A search is $5 per 1,000 searches of up to 10 results, and a page read is $0.80 per 1,000 pages. Neither needs a quote or an approval, so they come out of your balance up to the key's spending limit. Deep Search rows cost 5¢, 15¢ or $1, and a run can't cost more than the maximum approved. [Balance, quotes, and billing](https://specstream.ai/docs/billing.md) has the details.

## Remove

Delete `agent/connections/specstream.ts` and remove `SPECSTREAM_API_KEY` from the Vercel project. To stop the key working everywhere, revoke it under **API keys** in **Agents & API**. Results already delivered stay in your workspace. Questions go to [support@specstream.ai](mailto:support@specstream.ai).

## More

[SpecStream for eve](https://specstream.ai/integrations/eve.md) covers what it does in eve 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).
