All Installation Methods

Claude Code with MCP

claude mcp add futuresearch --scope project --transport http https://mcp.futuresearch.ai/mcp

Then launch Claude Code and authenticate with Google:

claude
/mcp → select FutureSearch → Authenticate

No API key needed. Try asking:

Which AI models had the biggest safety implications when released?

Full guide →

Claude Code with Plugin

The plugin bundles an MCP server and a skill that guides Claude to write Python SDK code for advanced pipelines.

claude plugin marketplace add futuresearch/futuresearch-python
claude plugin install futuresearch@futuresearch

Then launch Claude Code and authenticate:

claude
/mcp → select FutureSearch → Authenticate

No API key needed. Try asking:

Which AI models had the biggest safety implications when released?

Full guide →

Claude Cowork with MCP

Cowork is a tab in Claude Desktop for autonomous multi-step tasks.

1. Open Claude Desktop (download if needed)
2. Go to Settings → Connectors → Add custom connector
3. Enter the URL: https://mcp.futuresearch.ai/mcp
4. Go to Settings → Capabilities → Code execution and file creation → Additional allowed domains and add mcp.futuresearch.ai (this lets Claude upload your CSVs for the researchers to process)

Sign in with Google to authenticate. Switch to the Cowork tab and try:

Which S&P 500 companies are most exposed to China-Taiwan risk?

Full guide →

Python SDK with pip

pip install futuresearch

Requires an API key:

export FUTURESEARCH_API_KEY=sk-cho-...

Getting Started → | API Reference →

Dependencies: Python 3.12+ and pip.

Python SDK with uv

uv add futuresearch

Requires an API key:

export FUTURESEARCH_API_KEY=sk-cho-...

Getting Started → | API Reference →

Dependencies: Python 3.12+ and uv.

Codex with Skills

Ask Codex in natural language:

>>> codex
>>> Install the skill from github.com/futuresearch/futuresearch-python found under skills/futuresearch-python

Or install manually:

python ~/.codex/skills/.system/skill-installer/scripts/install-skill-from-github.py \
  --repo futuresearch/futuresearch-python --path skills/futuresearch-python

Restart Codex to pick up the new skill.

Important: be sure to supply your API key when launching Codex:

export FUTURESEARCH_API_KEY=sk-cho...
codex

Official Docs

Codex with MCP

Add FutureSearch to your Codex MCP config (requires uv):

[mcp_servers.futuresearch]
command = "uvx"
args = ["futuresearch-mcp"]

[mcp_servers.futuresearch.env]
FUTURESEARCH_API_KEY = "${FUTURESEARCH_API_KEY}"

Or use the CLI:

codex mcp add futuresearch --env FUTURESEARCH_API_KEY=${FUTURESEARCH_API_KEY} -- uvx futuresearch-mcp

Config file location:

• User scope: ~/.codex/config.toml
• Project scope: .codex/config.toml

Important: either insert your API key when creating the config file, or supply the key when launching Codex:

export FUTURESEARCH_API_KEY=sk-cho...
codex

Official Docs

Gemini with Skills

Install the skill directly from GitHub:

gemini skills install https://github.com/futuresearch/futuresearch-python --path skills/futuresearch-python

Use --scope workspace to install to .gemini/skills/ (project-level) instead of ~/.gemini/skills/ (user-level).

Then enable skills within the CLI:

/settings > Preview Features > Enable
/settings > Agent Skills > Enable
/skills enable futuresearch-python
/skills reload

Official Docs

Gemini with MCP

Add via CLI (requires uv):

gemini mcp add -e FUTURESEARCH_API_KEY=$FUTURESEARCH_API_KEY futuresearch uvx futuresearch-mcp

Or add manually to ~/.gemini/settings.json (user) or .gemini/settings.json (project):

{
  "mcpServers": {
    "futuresearch": {
      "command": "uvx",
      "args": ["futuresearch-mcp"],
      "env": {
        "FUTURESEARCH_API_KEY": "$FUTURESEARCH_API_KEY"
      }
    }
  }
}

Official Docs

Gemini with Plugin

This installs both the MCP server and skills together.

gemini extensions install https://github.com/futuresearch/futuresearch-python --ref main

When prompted with Would you like to attempt to install via "git clone" instead?, type Y to confirm.

Then enable the extension:

gemini extensions enable futuresearch [--scope <user or workspace>]

To use skills, enable them within the CLI:

/settings > Preview Features > Enable
/settings > Agent Skills > Enable
/skills enable futuresearch-python
/skills reload
/model > Manual > gemini-3-pro-preview

Official Docs

Cursor with Skills

1. Open Cursor Settings → Rules
2. In the Project Rules section, click Add Rule
3. Select Remote Rule (Github)
4. Enter: https://github.com/futuresearch/futuresearch-python.git

Official Docs

Cursor with MCP

Or add manually to your MCP config (requires uv):

{
  "mcpServers": {
    "futuresearch": {
      "command": "uvx",
      "args": ["futuresearch-mcp"],
      "env": {
        "FUTURESEARCH_API_KEY": "${env:FUTURESEARCH_API_KEY}"
      }
    }
  }
}

Config file location:

• macOS: ~/.cursor/mcp.json
• Windows: %USERPROFILE%\.cursor\mcp.json