JetBrains

Ask AI Assistant or Junie for live web data from your JetBrains IDE and get structured records back, with the same saved requests, results, projects, and balance as your SpecStream workspace.

SpecStream for JetBrainsCoding agents

What you need

A SpecStream account. New accounts get $10 of work on sign-up, no card.
A JetBrains IDE with AI Assistant or Junie.

1. Add the server in AI Assistant

Open Settings, then Tools, then AI Assistant, then Model Context Protocol (MCP), and select Add. Paste this JSON in the New MCP Server dialog.

Choose whether the server applies globally or to the current project, select OK, then select Apply.

New MCP Server
{
  "mcpServers": {
    "specstream": {
      "url": "https://api.specstream.ai/mcp"
    }
  }
}

2
2. Add the server in Junie
Open Settings, then Tools, then Junie, then MCP Settings, and select Add. In the mcp.json file that opens, add SpecStream under mcpServers with your API key as a header.
Create the key in the workspace: open Agents & API, choose API keys, then Create API key. Servers added in settings are saved to ~/.junie/mcp/mcp.json and apply to every project. Keep the key in that user file. A project's .junie/mcp/mcp.json can be committed.
~/.junie/mcp/mcp.json
{ "mcpServers": { "specstream": { "url": "https://api.specstream.ai/mcp", "headers": { "Authorization": "Bearer <your SpecStream API key>" } } } }

3. Ask for the data

Describe the records you want and the details to bring back on each one. For example:

In AI Assistant or Junie

Find Java libraries that generate PDFs, with each one's license, latest release, and Maven coordinates.

4. Or use the stdio proxy

For a local server, run pip install specstream-mcp-stdio and use this entry in either client. If the command is not on the IDE's PATH, put its absolute path in command.

mcp.json
{
  "mcpServers": {
    "specstream": {
      "command": "specstream-mcp-stdio",
      "args": [],
      "env": {
        "SPECSTREAM_API_KEY": "<your SpecStream API key>"
      }
    }
  }
}

AI Assistant's documentation lists Streamable HTTP, SSE, and stdio, and does not describe sign-in or headers for remote servers. If your IDE offers no way to sign in to SpecStream, use the stdio proxy with an API key.

Junie lists a remote server that needs OAuth with an Authorization required status. The Junie step above sends your API key as a header instead.

What happens when you ask

Ask in plain words, or ask the assistant to call compile with your request. SpecStream reads it back as editable parts, and quote_capacity_work prices it. Nothing is spent until you approve the maximum. authorize_capacity_work then starts the work, and rows arrive with the source URL and read time on every value.

To find saved work, ask the assistant to call list_specs or list_projects. To keep results up to date, ask for a schedule: the request reruns hourly to yearly inside your spending limits, and each rerun labels what changed. See schedules and the agent guide.

Cost

Reading a request and pricing it cost nothing. Each row that matches costs 5¢, 15¢ or $1, depending on how much the row needs, plus 5¢ for a person's email and 15¢ for a phone. A quote lasts 30 minutes, and a run that finds nothing costs nothing. How quotes and billing work

Disconnect

Remove SpecStream from the MCP settings in AI Assistant or Junie. On SpecStream's side, revoke the key in Agents & API, or revoke the IDE's access in Settings. Results you saved stay in your workspace. Questions go to [email protected].

SpecStream for JetBrains covers what it does in JetBrains on one page. Every other tool has its guide in Connect your tools. You see the most a request can cost before it starts, and the matches are paid from your balance. Billing has the details. For help, email [email protected].