Hyperdex

Apps and coding tools

Hyperdex works with any app or agent that supports the Model Context Protocol (MCP) over HTTP and can send an Authorization header. Replace hdx_your_key with your own key in every example.

Endpointhttps://hyperdex.hypervect.com/mcp
TransportStreamable HTTP (MCP 2026-07-28; 2025-era clients also work)
AuthenticationHeader Authorization: Bearer hdx_…

Claude Code

Run one command, then type /mcp inside Claude Code to check that Hyperdex is connected.

claude mcp add --transport http hyperdex https://hyperdex.hypervect.com/mcp \
  --header "Authorization: Bearer hdx_your_key"

Add --scope user to use Hyperdex in every project. In .mcp.json, each person sets their own HYPERDEX_API_KEY environment variable so no key is committed.

Claude (web and desktop app)

In Claude, open Customize, Connectors, Add custom connector and paste https://hyperdex.hypervect.com/mcp. For authentication choose No sign-in, then under Request headers add authorization with the value Bearer hdx_your_key. On Team and Enterprise plans an owner adds connectors in the organisation settings.

{
  "mcpServers": {
    "hyperdex": {
      "command": "npx",
      "args": ["-y", "mcp-remote", "https://hyperdex.hypervect.com/mcp",
               "--transport", "http-only", "--header", "Authorization:${AUTH_HEADER}"],
      "env": { "AUTH_HEADER": "Bearer hdx_your_key" }
    }
  }
}

Request headers on connectors are in beta and not yet available to every organisation. If you don’t see them, use the desktop fallback above with the community mcp-remote bridge (requires Node.js). The file is at ~/Library/Application Support/Claude/claude_desktop_config.json on macOS and %APPDATA%\Claude\claude_desktop_config.json on Windows; restart Claude after editing it.

Cursor

Add Hyperdex to .cursor/mcp.json in a project, or ~/.cursor/mcp.json for every project.

{
  "mcpServers": {
    "hyperdex": {
      "url": "https://hyperdex.hypervect.com/mcp",
      "headers": { "Authorization": "Bearer ${env:HYPERDEX_API_KEY}" }
    }
  }
}

VS Code (GitHub Copilot agent mode)

Add .vscode/mcp.json. VS Code asks for your key the first time and stores it securely.

{
  "inputs": [
    { "type": "promptString", "id": "hyperdex-key", "description": "Hyperdex API key", "password": true }
  ],
  "servers": {
    "hyperdex": {
      "type": "http",
      "url": "https://hyperdex.hypervect.com/mcp",
      "headers": { "Authorization": "Bearer ${input:hyperdex-key}" }
    }
  }
}

Note that VS Code uses a top-level servers key, not mcpServers.

Windsurf (Devin Desktop)

Edit ~/.config/devin/mcp_config.json on macOS and Linux, or %APPDATA%\devin\mcp_config.json on Windows.

{
  "mcpServers": {
    "hyperdex": {
      "serverUrl": "https://hyperdex.hypervect.com/mcp",
      "headers": { "Authorization": "Bearer ${env:HYPERDEX_API_KEY}" }
    }
  }
}

OpenAI Codex CLI

Codex reads the key from an environment variable and sends it as a bearer token.

export HYPERDEX_API_KEY=hdx_your_key
codex mcp add hyperdex --url https://hyperdex.hypervect.com/mcp --bearer-token-env-var HYPERDEX_API_KEY

Antigravity CLI and Gemini CLI

Google replaced Gemini CLI with Antigravity CLI for individual users. Antigravity reads ~/.gemini/config/mcp_config.json and accepts only serverUrl. Gemini CLI, still used with paid API keys, reads ~/.gemini/settings.json.

{
  "mcpServers": {
    "hyperdex": {
      "serverUrl": "https://hyperdex.hypervect.com/mcp",
      "headers": { "Authorization": "Bearer hdx_your_key" }
    }
  }
}

Cline

Open the MCP Servers icon, then Configure, and add Hyperdex to cline_mcp_settings.json. Set "type": "streamableHttp"; without it Cline falls back to an older transport.

{
  "mcpServers": {
    "hyperdex": {
      "type": "streamableHttp",
      "url": "https://hyperdex.hypervect.com/mcp",
      "headers": { "Authorization": "Bearer hdx_your_key" }
    }
  }
}

Continue

Save this as .continue/mcpServers/hyperdex.yaml and store your key as a Continue secret named HYPERDEX_API_KEY. MCP tools work in Agent mode.

name: Hyperdex
version: 0.0.1
schema: v1
mcpServers:
  - name: hyperdex
    type: streamable-http
    url: https://hyperdex.hypervect.com/mcp
    requestOptions:
      headers:
        Authorization: Bearer ${{ secrets.HYPERDEX_API_KEY }}

Zed

Add Hyperdex to context_servers in Zed’s settings.json.

{
  "context_servers": {
    "hyperdex": {
      "url": "https://hyperdex.hypervect.com/mcp",
      "headers": { "Authorization": "Bearer hdx_your_key" }
    }
  }
}

JetBrains (Junie and AI Assistant)

For the Junie CLI, add Hyperdex to ~/.junie/mcp/mcp.json (keep keys out of the project-level file, which can be committed). AI Assistant does not document custom headers, so connect it through the mcp-remote bridge (requires Node.js) in Settings, Tools, AI Assistant, Model Context Protocol.

{
  "mcpServers": {
    "hyperdex": {
      "url": "https://hyperdex.hypervect.com/mcp",
      "headers": { "Authorization": "Bearer hdx_your_key" }
    }
  }
}

GitHub Copilot cloud agent

In the repository, open Settings, Copilot, MCP servers and paste this. Add your key as an Agents secret named COPILOT_MCP_HYPERDEX_API_KEY. GitHub recommends listing only the tools the agent needs instead of *.

{
  "mcpServers": {
    "hyperdex": {
      "type": "http",
      "url": "https://hyperdex.hypervect.com/mcp",
      "headers": { "Authorization": "Bearer $COPILOT_MCP_HYPERDEX_API_KEY" },
      "tools": ["search", "get_briefing", "log_activity", "list_tasks"]
    }
  }
}

OpenAI API (Responses and Agents SDK)

Give the hosted mcp tool your key in headers. OpenAI calls Hyperdex from its own servers. By default it asks for approval before each tool call; set require_approval to change that.

from openai import OpenAI

client = OpenAI()
response = client.responses.create(
    model="gpt-6-astra",
    input="Prep me for my call with Jane at Acme.",
    tools=[{
        "type": "mcp",
        "server_label": "hyperdex",
        "server_url": "https://hyperdex.hypervect.com/mcp",
        "headers": {"Authorization": "Bearer hdx_your_key"},
        "require_approval": "never",
    }],
)
print(response.output_text)

Anthropic API (MCP connector)

Pass Hyperdex in mcp_servers with your key as authorization_token, using the MCP connector beta.

import anthropic

client = anthropic.Anthropic()
response = client.beta.messages.create(
    model="claude-opus-5-5",
    max_tokens=1024,
    messages=[{"role": "user", "content": "Prep me for my call with Jane at Acme."}],
    mcp_servers=[{
        "type": "url",
        "url": "https://hyperdex.hypervect.com/mcp",
        "name": "hyperdex",
        "authorization_token": "hdx_your_key",
    }],
    tools=[{"type": "mcp_toolset", "mcp_server_name": "hyperdex"}],
    betas=["mcp-client-2026-09-15"],
)

ChatGPT

ChatGPT connectors cannot send an API key today: they support only OAuth sign-in or no authentication. Hyperdex uses API keys, so ChatGPT cannot connect yet. You can use the OpenAI API above instead.

Troubleshooting

Checked against each vendor’s own documentation in October 2026. Products change quickly; if a step no longer matches, write to transform@hypervect.com and we will update it.