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

# VoidLLM

Register SpecStream as an external MCP server in VoidLLM's Admin UI, and every client behind your gateway can search the live web, read pages and ask for structured data through one VoidLLM endpoint, on one SpecStream key with a spending limit.

[SpecStream for VoidLLMAgent frameworks & builders](https://specstream.ai/integrations/voidllm.md)

## What you need

-   A VoidLLM install where you are a team admin, org admin or system admin, and a VoidLLM API key (`vl_uk_...`) to test with.
-   A SpecStream account. New accounts start with $10 of credit, no card. [Create one](https://specstream.ai/signup?utm_source=voidllm).

1.  1

    ## 1. Create a SpecStream API key

    In SpecStream, open [Agents & API](https://specstream.ai/workspace/agents?tab=keys&utm_source=voidllm), choose **API keys**, then **Create API key**, and give it a total spending limit above 0.

    VoidLLM sends one token per server to the upstream, and its form warns that credentials configured there are shared across all users with access to the server. Every user of the server spends from this key, so its limit is the ceiling for all of them. Make a key for the gateway alone, not your personal one.

2.  2

    ## 2. Add the server in the Admin UI

    Open **MCP Servers** in the sidebar and click **Add Server**. Fill in the **Add MCP Server** dialog:

    **Scope**: **Team**, **Organization** or **Global**, as your role allows. A team or organization server is open to its members at once. A global server needs an access grant (below).

    **Name**: `SpecStream`. **Alias**: `specstream`. The alias names the server's endpoint and its tools in Code Mode, and must be unique.

    **URL**: `https://api.specstream.ai/mcp`. SpecStream speaks Streamable HTTP, the transport VoidLLM proxies.

    **Auth Type**: **Bearer Token**. **Auth Token**: the SpecStream key. VoidLLM sends it upstream as `Authorization: Bearer <key>` and stores it encrypted.

    Leave **Protocol Version** at **Auto-detect (recommended)** and click **Add Server**.

3.  3

    ## 3. Test it and look at the tools

    In the server's row, click **Test connection**. Then click the chevron to open the server: its tools are listed there, and **Refresh Tools** reads the list again. You should see `search_web`, `read_pages` and the Deep Search tools such as `compile` and `quote_capacity_work`.

4.  4

    ## 4. Grant access to a global server

    Skip this for a team or organization server. A global server is closed by default: go to **Organization**, open the **MCP Servers** tab, turn SpecStream on and save. A team can narrow what its organization allows, never widen it.

5.  5

    ## 5. Connect a client

    Each registered server gets its own endpoint, `/api/v1/mcp/specstream`. Open the server and click **Copy** for the client config, or add this entry to the client's `mcpServers`, with a VoidLLM key in place of the placeholder:

    MCP client config

    ```json
    {
      "mcpServers": {
        "specstream": {
          "url": "https://your-voidllm/api/v1/mcp/specstream",
          "headers": {
            "Authorization": "Bearer vl_uk_..."
          }
        }
      }
    }
    ```

## Ask for something from this week

In the client, ask a question with a recent answer. The model calls `search_web` with `{"q": "<the question>"}` and gets back an `answer` with the page it was read from, and the `sources` it read, each with `url`, `title`, `site` and `snippet`. To read a page it found, it calls `read_pages` with `{"urls": ["<url>"]}` and gets the page's main text as Markdown. VoidLLM logs the server, tool, duration and status of each call, not its arguments or results.

## Or register it by API or YAML

With the Admin API, a key that may manage servers posts the same fields:

## In the config file

Or add it under `mcp_servers` in VoidLLM's YAML config, with the key in `SPECSTREAM_API_KEY` in VoidLLM's environment. A YAML server is synced to the database at startup, so a change takes a restart:

## Use it in Code Mode

Code Mode gives a model one `execute_code` tool that runs JavaScript, with each server's tools as async functions under its alias. A search and a page read in one run:

## Keep it to search and reading

Deep Search tools start paid runs only after a quote: `authorize_capacity_work` needs the approved maximum in `confirm_max_cents`, and a run cannot cost more. To keep the gateway's users to the Search API and Extract API alone, open the server's tools and turn off `authorize_capacity_work` in VoidLLM's tool blocklist. VoidLLM's server docs say a blocked tool returns an error when called directly, and its Code Mode page says a blocked tool still works through the regular proxy, so call it once through the gateway to see which holds in your version.

## Cost

Registering SpecStream in VoidLLM 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` reads is $0.80 per 1,000 pages; a page with no text is not charged. Both come out of the balance with no quote and no approval, within the key's spending limit. Deep Search is priced per row, and nothing runs until its quote is approved. [Search API](https://specstream.ai/search-api.md), [Extract API](https://specstream.ai/extract-api.md) and [billing](https://specstream.ai/docs/billing.md).

## Remove

Delete SpecStream on VoidLLM's **MCP Servers** page, or take it out of the YAML file and restart. Revoke the key in SpecStream's **Agents & API** to end its access at once. Every tool on the server is described in the [agent guide](https://specstream.ai/docs/mcp.md). Questions go to [support@specstream.ai](mailto:support@specstream.ai).

## More

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