Claude, ChatGPT, Cursor, VS Code and Windsurf

Use IMPACT in Claude, ChatGPT, Cursor, VS Code and Windsurf

For marketing, sales and customer success teams. The decision: whether to run IMPACT in a browser with no sign in, or inside the assistant you already use.

You need one address and no sign in. Pick your assistant and follow its steps.

Where it works: in a web browser with no sign in, in Claude as a connector (including Claude Code and Cowork), in ChatGPT, in Gemini through the Helix GTM Suite skills package, and in Cursor, VS Code and Windsurf as an MCP server.

By Shashwat Ghosh, Co-Founder and Fractional CMO, Helix GTM Consulting · Last updated 5 November 2026

How it looks in Claude

Ask in plain words and Claude calls the connector. The example uses ICP Intelligence; every connector works the same way.

All six: 12,400 npm downloads, first release to 4 November 2026 (GTM Alpha first released 8 September 2025, the other five in January 2026). npm also counts automated downloads.

Free to install. Seven are published in Anthropic's plugin directory; B2B ICP and Customer Insight is in review; B2B Deal Execution is public on GitHub and installable. With the 48 connector tools in six connectors, that makes 85 free tools and skills. Each plugin page shows real answers from a test run.

A real answer from the ICP Intelligence connector. The question gave no match rate or share, so the tool used labelled examples.

Quick start: three steps

Add this MCP server address to Claude, ChatGPT, Cursor, VS Code or Windsurf: https://impact.gtmhelix.com/mcp
Leave the sign in fields empty. The steps for each assistant are below.
Ask in plain words, for example: Use IMPACT to identify likely champions for our product, cloud security monitoring that ranks misconfigurations by real exposure, for mid-size fintech and SaaS companies.

Try these prompts

Use IMPACT to identify likely champions for our product, cloud security monitoring that ranks misconfigurations by real exposure, for mid-size fintech and SaaS companies.
Map the alternatives for our product in the cloud security monitoring category, compared with your first competitor and your second competitor.
Run an IMPACT full audit of our positioning: cloud security monitoring for companies on AWS, Azure and Google Cloud, target customer CISOs at mid-size fintech and SaaS companies, problem solved too many cloud alerts with no clear order to fix them.

When a tool is not enough: if the decision needs more than one answer, talk it through in a call. Book 45 minutes with Shashwat.

Before you connect

Claude plan: any. Custom connectors work on Claude's Free, Pro, Max, Team and Enterprise plans; the Free plan allows one custom connector.
Team or Enterprise: only an Owner can add a custom connector, under Organization settings, Connectors, Add, Custom, Web. Members then turn it on under Customize, Connectors.
Accounts and keys: none. There is no sign in and no API key.
Local use: the npm package needs Node.js 18 or later.

What IMPACT does and does not do

All 8 tools are read only: each one builds a text answer from the inputs you give it and changes nothing anywhere. What you type is used once to build the answer and is not stored or logged. It is not ideal for:

Teams that need buyer research done for them: IMPACT structures your positioning as hypotheses, and you test them with buyers.
Anyone who needs live competitor data: the tools do not look anything up. They work from what you type about your market and rivals, and anything they add themselves (such as common complaints in a category) is marked as something to check with buyers.
Teams that need a saved history: answers are not stored, so save each draft wherever you keep your positioning work.
Teams that need IMPACT to write to their CRM or other tools: none of the tools can.

The MCP server address

https://impact.gtmhelix.com/mcp

It uses Streamable HTTP, is stateless and needs no sign in. All 8 tools are read only. The older address https://impact-mcp.netlify.app/mcp still answers directly for anyone who set it up earlier.

Tools:

IMPACT Framework Guide: Helps you decide which positioning steps to run, and in what order. You give nothing, or one IMPACT step to focus on and your sector. You get the IMPACT positioning method step by step (identify champions, map alternatives, pinpoint value, anchor in a market, craft the message, translate it to execution), with a template or checklist for each step; a focus step returns that step only, and your sector adds its buyer roles, measures and proof points.
Identify Champions: Helps you decide who inside the buyer will push for you. You give what your product does and the problem it solves, plus optional company name, target company type, price point and sector. You get hypotheses for the champion, the economic buyer, the technical influencer, commercial and control reviewers and daily users, warning signs of an anti champion, a path to develop the champion and what to give them. These are roles to test, not named people.
Map Alternatives: Helps you decide what buyers compare you with, and where you can lead. You give your product and its category, plus optional competitors, their weaknesses and your strengths. You get one card for each competitor you name, the status quo and doing nothing, where you can lead and discovery questions, with your sector's questions and objections. It uses only what you give and says what would sharpen the map.
Pinpoint Value: Helps you decide what you can claim that buyers will believe. You give your target customer, the main outcome they get and what you do that others do not, plus optional product name, category, customer results and business model. You get value statements in three versions, a value matrix for your own figures, proof points ranked from customer results to social proof, versions for different audiences and channels, and questions to test the value.
Anchor Market: Helps you decide which market segment to win first. You give what your product does and the segments you are weighing, plus optional customers, pain, deal size and company counts. You get a beachhead choice: your segments ranked from your own customers, pain and deal size when you give all three (the answer says which method it used), otherwise scored on keywords in their names; a TAM, SAM and SOM calculation from only the figures you give; a three year expansion path and an ICP hypothesis for the chosen segment. It never chooses a segment you did not name.
Craft Message: Helps you decide what your core message is, and how to test it. You give your target customer, the key benefit and what sets you apart, plus optional product name, customer need, category, main competitor and business model. You get a positioning statement (full and one sentence), a message hierarchy from tagline to three supporting pillars, versions for different audiences, objection handling and ways to test the message. It writes from your own words, in whole sentences.
Translate Execution: Helps you decide how your positioning should read in each channel. You give a positioning statement, your target customer and the key benefit, plus optional channels, product name and business model. You get copy to edit: a website hero, a LinkedIn tagline and post drafts, a three email cold sequence, a sales deck outline and a 15 minute demo script, with calls to action that fit your business model and the order to work the channels in.
IMPACT Full Audit: Helps you decide whether your positioning is working and what to fix first. You give what your product does, your target customer and the problem you solve, plus optional company name, differentiation, competitors, current positioning, customer feedback and business model. You get a positioning audit: an input completeness score for each IMPACT step, the proof you already have, your current positioning checked in words, a recommended positioning with tagline options, a 30 day action plan and the measures to track. The score measures how complete your inputs are, not how strong your positioning is.

Support: shashwat@gtmhelix.com or GitHub issues

Limits: up to 262,144 bytes per call, text fields of up to 4,000 characters and one call per request (no batches). A call over a limit is refused before the tool runs.

Claude (web, desktop and mobile)

Open Customize, then Connectors, select +, then Add custom connector.
Name: IMPACT. Remote MCP server URL: https://impact.gtmhelix.com/mcp.
Leave the sign in fields empty and select Add.
In a chat, turn the connector on from the tools menu and ask, for example: Use IMPACT to identify likely champions for our product, cloud security monitoring that ranks misconfigurations by real exposure, for mid-size fintech and SaaS companies.

On Team and Enterprise, an Owner adds the connector first under Organization settings, Connectors, Add, Custom, Web; each member then turns it on under Customize, Connectors.

Claude Code

Terminal

claude mcp add --transport http impact https://impact.gtmhelix.com/mcp

ChatGPT

ChatGPT can use remote MCP servers as custom connectors on plans that allow developer mode. In ChatGPT's settings, open the connectors (apps) section, turn on developer mode if your plan requires it, create a connector with the URL https://impact.gtmhelix.com/mcp and no authentication, then enable it in a chat. Menu names change between ChatGPT releases; OpenAI's help centre has the current path.

Cursor

Add this to ~/.cursor/mcp.json (all projects) or .cursor/mcp.json (one project):

~/.cursor/mcp.json

{
  "mcpServers": {
    "impact": {
      "url": "https://impact.gtmhelix.com/mcp"
    }
  }
}

VS Code (GitHub Copilot agent mode)

Add this to .vscode/mcp.json in your project:

.vscode/mcp.json

{
  "servers": {
    "impact": {
      "type": "http",
      "url": "https://impact.gtmhelix.com/mcp"
    }
  }
}

Windsurf

Add this to ~/.codeium/windsurf/mcp_config.json:

~/.codeium/windsurf/mcp_config.json

{
  "mcpServers": {
    "impact": {
      "serverUrl": "https://impact.gtmhelix.com/mcp"
    }
  }
}

Other assistants, or no internet access

Any MCP client that supports Streamable HTTP can use the address above with no authentication. For a client that only runs local servers, use the npm package, which runs these tools on your own computer over stdio (npm can lag behind this site until a new version is published):

MCP client config (npm package)

{
  "mcpServers": {
    "impact": {
      "command": "npx",
      "args": [
        "-y",
        "@shashwatgtmalpha/impact-mcp"
      ]
    }
  }
}

How to read the answer

When your assistant calls an IMPACT tool, the server returns a structured Markdown positioning analysis or messaging draft built from fixed rules, with bracketed placeholders for missing facts and labelled example figures for any benchmark you did not supply.

What to do next

Founder: Add the connector URL in your assistant and run Full Audit or Identify Champions with your real product and buyer details.
CEO: Share this setup page with your marketing and sales leads so everyone works from the same positioning rubric.
CMO: Connect IMPACT in your assistant to run the six positioning steps from champion definition through channel translation.
CRO: Use the connected Map Alternatives and Translate Execution tools to sharpen competitive positioning and sales talk tracks.
Demand generation head: Run Craft Message and Translate Execution in your assistant to generate aligned headlines and campaign copy.

Why come back

Come back when you add a new assistant, onboard a teammate, or check the example prompts and payload limits. The tool applies the same fixed rules every time. It does not call an AI model, and this site does not keep what you enter.

Choose your app

Run a Helix tool in your browser, then take the result to ChatGPT, Claude or Gemini. Connectors work with supported plans in Claude, ChatGPT, Cursor, VS Code and Windsurf. The Helix GTM Suite skills package supports Claude Code, Gemini CLI, Antigravity, Cursor and VS Code, and is available through skills.sh. Each plugin has its own installation steps, below or on its own page.

Connector setup guides · Skills package setup guide · Find the skills on skills.sh