Build with
GovCon MCP.

Build a workflow

Begin with the member's task, then inspect the tools offered by the authenticated connection. Use the tool reference to choose a discovery, detail, document or account-action step before constructing its arguments.

Carry returned identifiers between steps. A company search result, a public opportunity and a CRM account are different destinations; resolve the intended record before reading more detail or saving work.

Keep reading and analysis as separate choices. Retrieve existing documents and results first, then start G2X-hosted analysis only when the member requests it. Continue an accepted analysis with its returned run identifier.

Quickstart

You need a G2X account; review G2X plans.

1. Add https://mcp.g2x.com/mcp to your agent as a remote MCP server. Find your agent's steps.
2. Sign in with your G2X account when your agent asks.
3. Ask “Look up [Company Name or UEI] in G2X.”

Installing the agent skill does not connect or authorize your account.

Add the GovCon MCP skill for guidance on choosing tools, keeping citations and completing account workflows.

Sign in before tool discovery: tools/list and every data or tool call require an authenticated G2X account. The transport is stateless, so do not depend on a server session ID.

Official integrations

G2X's own listings on other platforms.

G2X in ChatGPT

G2X's official plugin in ChatGPT, published in ChatGPT's plugin directory.

1. Open G2X in ChatGPT's plugin directory.
2. Choose Install plugin.
3. Sign in with your G2X account when ChatGPT asks.

A G2X account is required. Every G2X plan works, including the free Community Edition.

G2X in ChatGPT's plugin directory

G2X GovCon Growth OS (Grok Bot)

G2X's own bot, published on xAI's Grok Bot.

1. Open G2X GovCon Growth OS on x.ai.
2. Choose Add to Grok Bot, then finish in the Grok Bot app.
3. When Grok Bot prompts you to connect G2X, sign in with your G2X account to approve.

G2X GovCon Growth OS on x.ai

G2X on Vercel Connect

G2X's connector on Vercel Connect, for the apps and agents you deploy on Vercel.

1. Open G2X on Vercel Connect and choose Create Connector, or run vercel connect create g2x in your linked project.
2. In your app, request a G2X token for each signed-in user with the Vercel Connect SDK. The first time, send the user to sign in with their G2X account.
3. Use the token with your MCP client at https://mcp.g2x.com/mcp.

Each user needs a G2X account. Every G2X plan works, and each user's plan sets what G2X returns.

G2X on Vercel Connect

Connect via MCP

Every agent that supports MCP connects to G2X: add https://mcp.g2x.com/mcp and sign in with your G2X account. Pick your agent for its steps.

Tested: a working G2X connection is on record. Steps checked: the steps follow the agent's own guide.

ChatGPT

Steps checked

ChatGPT is an AI agent. Install the G2X plugin from ChatGPT's plugin directory, or add G2X yourself as a custom MCP server.

The quickest way: G2X's official plugin, published in ChatGPT's plugin directory.

1. Open G2X in ChatGPT's plugin directory.
2. Choose Install plugin.
3. Sign in with your G2X account when ChatGPT asks.

A G2X account is required. Every G2X plan works, including the free Community Edition. What ChatGPT can see and do follows your G2X account's permissions.

Or add G2X yourself as a custom MCP server:

1. In ChatGPT on the web, go to chatgpt.com/plugins, choose +, then Add custom MCP server. If ChatGPT asks, turn on Developer mode first.
2. Enter G2X as the name. Under Connection, enter https://mcp.g2x.com/mcp as the Server URL. For Authentication, choose the sign-in option.
3. Read the warning, choose I understand and want to continue, then choose Create as a plugin.
4. Open your personal plugins at chatgpt.com/plugins?view=personal, open G2X and choose + to install it. If you created it in a workspace, it's in that workspace's plugins.
5. Sign in with your G2X account when ChatGPT asks.

In a Business, Enterprise or Edu workspace, your admin turns on plugins with MCP servers, and then Add custom MCP server appears.

Steps checked against G2X in ChatGPT's plugin directory and the ChatGPT guide to connecting a server.

ChatGPT setup page

Claude

Tested

Claude is Anthropic's AI agent. You add G2X once as a custom connector.

1. In Claude, open Customize → Connectors.
2. Choose + Add, then Add custom connector.
3. Enter G2X as the name and https://mcp.g2x.com/mcp as the server URL, then choose Continue until Claude shows its sign-in settings.
4. Keep Sign in now, choose Register automatically, then choose Add. If Claude shows Advanced settings instead, leave the client ID and client secret blank and choose Add.
5. If Claude shows Connect, choose it. Sign in with your G2X account.
6. Choose Allow to connect your account to Claude.
7. In a new chat, choose +, then Connectors, and turn on G2X. Ask “Look up [Company Name or UEI] in G2X.” If Claude asks to use a G2X tool, review the request and choose Allow once.

On the Free plan you can add one custom connector. On a Team or Enterprise plan, an owner adds G2X once in Organization settings → Connectors (Add → Custom → Web), with the same name, address and sign-in settings. Everyone else then opens Customize → Connectors, finds G2X and chooses Connect.

Claude setup page

Grok Bot

Steps checked

Grok Bot is an AI agent on the web and on iOS and Android. You add G2X once as a custom connector.

1. Go to grok.com/connectors. In the mobile app, open Settings → Connectors.
2. Choose New Connector, then Custom.
3. Paste https://mcp.g2x.com/mcp and sign in with your G2X account.

On a business or enterprise account, an admin sets up the connector for your team first.

Steps checked against the Grok Bot connectors guide.

Grok Bot setup page

Microsoft Copilot

Steps checked

Microsoft Copilot uses G2X through Microsoft 365, with a work or school account. A Copilot Studio maker adds G2X to an agent, and people in your organization use that agent in Microsoft Copilot.

Add G2X to a Copilot Studio agent

1. In Copilot Studio, open your agent and go to Tools. Choose Add a tool → New tool → Model Context Protocol.
2. Enter G2X as the server name, a short description, and https://mcp.g2x.com/mcp as the server URL.
3. For authentication, choose the sign-in option, then Dynamic discovery. Choose Create.
4. Choose Create a new connection, sign in with your G2X account, then choose Add to agent.
5. Publish the agent to the Teams and Microsoft 365 channel, so people can use it in Microsoft Copilot.

Microsoft 365 admins can also add MCP servers for the whole organization as connectors in the Microsoft 365 admin center. The Copilot app for a personal Microsoft account connects only to a fixed list of services, such as OneDrive and Gmail. It can't add G2X.

Steps checked against Microsoft's guide to adding an MCP server to an agent and Microsoft's guide to connecting Copilot to other services.

Microsoft Copilot setup page

Gemini

Steps checked

Gemini is Google's AI agent. Add G2X to the Gemini app on the web, or to Gemini CLI in your terminal.

Gemini app

1. On gemini.google.com, open Settings, then Connected Apps.
2. Under Custom apps, choose Add a custom app and paste https://mcp.g2x.com/mcp.
3. Choose Next, then sign in with your G2X account.

Google offers custom apps to personal Google Accounts of people 18 or older in the US, using Gemini in English with Keep Activity on.

Gemini CLI

gemini mcp add --transport http govcon https://mcp.g2x.com/mcp

1. Gemini CLI finds the G2X sign-in on its own. Sign in with your G2X account when it asks.

Steps checked against Google's help on custom apps in Gemini and the Gemini CLI guide to MCP servers.

Gemini setup page

Notion

Steps checked

Notion's AI agents can use G2X on Notion's Business and Enterprise plans. Add G2X to your Notion Agent, or to a Custom Agent your team shares.

Notion Agent

1. In Notion, open Settings → Connections and choose the Discover tab.
2. Choose Add Custom MCP and paste https://mcp.g2x.com/mcp.
3. Sign in with your G2X account when Notion asks.

Custom Agents

1. Open the agent's Settings and go to Tools & Access.
2. Choose Add connection → Custom MCP server.
3. Paste https://mcp.g2x.com/mcp as the server URL, enter G2X as the display name, then choose Connect.
4. Sign in with your G2X account when Notion asks.

A workspace admin turns on custom MCP servers first, under Settings → Connections → Manage.

Steps checked against Notion's guide to MCP servers for your Notion Agent and Notion's guide to MCP connections for Custom Agents.

Notion setup page

Muse

Steps checked

Muse Code is Meta's AI coding agent. You add G2X to its settings file, then sign in.

Muse Code

Add G2X to the mcp_servers block of ~/.config/muse/settings.json:

{
  "mcp_servers": {
    "govcon": {
      "transport": "streamable_http",
      "url": "https://mcp.g2x.com/mcp",
      "enabled": true
    }
  }
}

1. Run muse mcp login govcon and sign in with your G2X account in your browser.

The G2X connector for the Muse app is anticipated to be live soon.

Steps checked against Muse Code's guide to MCP servers.

Muse setup page

Cursor

Steps checked

Cursor is an AI code editor. You can add G2X in one click.

Add G2X to Cursor

1. Choose Add G2X to Cursor. Cursor opens to add the G2X server, named govcon.
2. When Cursor opens your browser, sign in with your G2X account.

Add it by hand

Add G2X to ~/.cursor/mcp.json for all your projects, or to .cursor/mcp.json in one project:

{
  "mcpServers": {
    "govcon": {
      "url": "https://mcp.g2x.com/mcp"
    }
  }
}

Steps checked against Cursor's guide to MCP install links and Cursor's MCP guide.

Cursor setup page

Zapier

Steps checked

Zapier connects your apps in automated workflows called Zaps. Its MCP Client app, in beta, connects G2X.

1. In Zapier, open Apps, choose Add connection, and search for MCP Client.
2. Paste https://mcp.g2x.com/mcp as the server URL and choose Streamable HTTP.
3. For the sign-in setting, choose Yes.
4. Choose Continue and sign in with your G2X account.
5. Use MCP Client actions in your Zaps to call G2X.

Steps checked against Zapier's MCP Client guide.

Zapier setup page

Devin

Steps checked

Devin is an AI software engineer. An organization admin adds G2X once for the team.

Devin

1. Open Customize → MCPs and choose Add custom MCP.
2. Enter G2X as the name, choose HTTP as the transport, and paste https://mcp.g2x.com/mcp.
3. For authentication, choose the sign-in option, for the whole organization or for each person. Then sign in with your G2X account.

Devin CLI

devin mcp add govcon https://mcp.g2x.com/mcp
devin mcp login govcon

1. Sign in with your G2X account when Devin CLI asks.

Devin Desktop

Devin Desktop, formerly Windsurf, reads its servers from an MCP config file.

1. In the Cascade panel, open the … menu and choose Open MCP config file.
2. Add a server named govcon with serverUrl set to https://mcp.g2x.com/mcp, then save the file.
3. Sign in with your G2X account when Devin Desktop asks.

Steps checked against Devin's MCP guide, the Devin CLI MCP guide and the Devin Desktop MCP guide.

Devin setup page

Claude Code

Steps checked

Claude Code is Anthropic's AI coding agent for your terminal. You add G2X with one command.

claude mcp add --transport http govcon https://mcp.g2x.com/mcp

1. In Claude Code, type /mcp, choose govcon and sign in with your G2X account in your browser.

Steps checked against the Claude Code guide to MCP.

Claude Code setup page

VS Code

Steps checked

VS Code is Microsoft's code editor. Its chat, including GitHub Copilot, can use G2X once you add the G2X server. You can add it in one click.

Install G2X in VS Code

1. Choose Install G2X in VS Code. Or run MCP: Add Server from the Command Palette, choose HTTP, paste https://mcp.g2x.com/mcp and name it govcon.
2. When VS Code opens your browser, sign in with your G2X account.

If your organization uses GitHub Copilot Business or Enterprise, an admin turns on the MCP servers in Copilot policy first.

Steps checked against VS Code's MCP guide and GitHub's guide to MCP in Copilot Chat.

VS Code setup page

OpenAI Codex

Steps checked

Codex is an AI coding agent. The Codex CLI, the Codex IDE extension and the ChatGPT desktop app share one setup, so you add G2X once.

Codex CLI

codex mcp add govcon --url https://mcp.g2x.com/mcp
codex mcp login govcon

1. codex mcp login govcon opens your browser. Sign in with your G2X account.

Codex in the app or your editor

1. In the ChatGPT desktop app or the Codex IDE extension, open Settings → MCP servers → Add server.
2. Choose Streamable HTTP and paste https://mcp.g2x.com/mcp.
3. Choose Authenticate and sign in with your G2X account, then restart the app.

Steps checked against the Codex guide to MCP.

OpenAI Codex setup page

Perplexity

Steps checked

Perplexity is an AI answer engine. You add G2X once as a custom remote connector.

1. Open Account settings → Connectors and choose + Custom connector.
2. Choose Remote.
3. Enter G2X as the name and https://mcp.g2x.com/mcp as the MCP server URL.
4. For Authentication, choose the sign-in option. For Transport, choose Streamable HTTP.
5. Check the acknowledgement box and choose Add.
6. Choose the G2X card to sign in with your G2X account and turn the connector on.

On Perplexity Enterprise, members add custom connectors after an admin allows it.

Steps checked against Perplexity's guide to custom remote connectors.

Perplexity setup page

Le Chat

Steps checked

Le Chat is Mistral AI's agent. An admin adds custom connectors. On Free, Pro and Student plans, the account owner is the admin.

1. Open the Connectors page and choose + Add Connector.
2. Choose the Custom MCP Connector tab.
3. Enter g2x as the name, with no spaces, and paste https://mcp.g2x.com/mcp.
4. Choose Connect. Le Chat finds the G2X sign-in on its own.

Steps checked against Mistral's guide to MCP connectors in Le Chat.

Le Chat setup page

Manus

Steps checked

Manus is an AI agent that works through tasks for you. You add G2X once as a custom MCP server.

1. Open Settings → Integrations → Custom MCP Servers and choose Add Server.
2. Enter G2X as the name, paste https://mcp.g2x.com/mcp, and add the server.

Steps checked against Manus's guide to custom MCP servers.

Manus setup page

Factory

Steps checked

Droid is Factory's AI coding agent. You add G2X to it from your terminal.

droid mcp add govcon https://mcp.g2x.com/mcp --type http

1. In Droid, type /mcp and choose govcon.
2. Sign in with your G2X account when your browser opens.

Steps checked against Factory's MCP guide.

Factory setup page

n8n

Steps checked

n8n is a workflow automation tool. Its MCP Client nodes connect G2X to a workflow or to an AI Agent.

1. Add an MCP Client node to your workflow, or an MCP Client Tool node to an AI Agent.
2. Choose the streamable HTTP transport and paste https://mcp.g2x.com/mcp as the MCP endpoint URL.
3. For Authentication, create an MCP sign-in credential with Dynamic Client Registration turned on.
4. Sign in with your G2X account.

Steps checked against n8n's MCP Client node guide and n8n's MCP credentials guide.

n8n setup page

Make

Steps checked

Make is a visual automation platform. Its MCP Client module connects G2X to a scenario.

1. Add the MCP Client module to a scenario and choose Create a connection.
2. Choose + New MCP Server and paste https://mcp.g2x.com/mcp.
3. If Make asks how to sign in, choose the sign-in option. Save the connection.
4. Sign in with your G2X account.

Steps checked against Make's MCP Client guide.

Make setup page

Replit

Steps checked

Replit Agent builds apps for you. You add G2X once as an MCP server.

Add G2X to Replit

1. Choose Add G2X to Replit, finish the install in Replit, and sign in with your G2X account when Replit asks.
2. Or go to replit.com/integrations, scroll to MCP Servers for Replit Agent and choose + Add MCP server. Enter G2X as the name, paste https://mcp.g2x.com/mcp, then choose Test & save and sign in with your G2X account.

Steps checked against Replit's guide to connecting via MCP and Replit's guide to MCP install links.

Replit setup page

Hermes Agent

Steps checked

Hermes Agent is an open-source AI agent from Nous Research.

Option A: commands

Run these in your terminal:

hermes config set mcp_servers.govcon.url https://mcp.g2x.com/mcp
hermes config set mcp_servers.govcon.auth oauth
hermes config set mcp_servers.govcon.oauth.cimd false
hermes mcp login govcon

The last command opens your browser so you can sign in.

Option B: edit config.yaml

Open ~/.hermes/config.yaml and add G2X under mcp_servers:

mcp_servers:
  govcon:
    url: "https://mcp.g2x.com/mcp"
    auth: oauth
    oauth:
      cimd: false

1. Save the file. Then, in a new terminal window, run hermes mcp login govcon.

Steps checked against Hermes Agent's MCP guide and MCP config reference.

Hermes Agent setup page

Any agent via MCP

Other agents need to support remote MCP servers with sign-in. Add https://mcp.g2x.com/mcp as a remote MCP server, then sign in with your G2X account when your agent asks.

Authentication

Use G2X OAuth in clients that support it. Without G2X sign-in, tools/list and every tool call return HTTP 401 with a WWW-Authenticate header that links the OAuth protected-resource metadata. Follow the advertised authorization server's authorization-code flow with PKCE S256.

Claude connects with the server URL alone. Every agent under Connect via MCP follows its own steps to the same sign-in. A G2X API key is not a replacement for this sign-in flow. Keep credentials in the client's connection settings, never in a URL or a prompt.

Every tool requires authentication. Your account determines access and allowances. Signing in never grants access to another account's watchlist, workspace or analysis.

Tool catalog

MCP tools/list is the authoritative source for input schemas, descriptions and annotations. Read the schema before constructing arguments. Record IDs must come from search results.

Every tool, with inputs and examples

Document formats and completeness

Ask for the complete document in Markdown, JSON or both. Markdown is for readable text; JSON preserves the supplied document structure. A page, excerpt or list of document names does not satisfy that request. Keep the document version, tables and citations with the result.

If your connection cannot return a complete artifact, use a supported source-document link. Do not present partial text as the full document or start analysis just to obtain it.

Capture workflow

1. Find an opportunity and keep its record ID.
2. Retrieve stored document content and existing analysis first. Keep each document's version and citations, and distinguish complete artifacts from excerpts.
3. Use the user's chosen agent for new interpretation, qualification and drafting by default. Save requested outputs through supported G2X actions. Do not start hosted analysis merely to read documents or save work.
4. Only if the user explicitly chooses G2X-hosted analysis, select the documents and use g2x_shred_opportunity for requirements or g2x_qualify_opportunity for company fit. Follow the connected tool schemas.
5. For a hosted run, poll g2x_get_analysis_run with the returned run ID. An accepted run is not a completed result. Retrieve the compliance matrix with the qualification run ID when its schema requires that identifier.

Make only the account changes the user asked for, and resolve an ambiguous target before acting. If a response is slow, resume the run by its existing ID instead of starting another.

Plans and usage

G2X pricing is the source for G2X plans and features. Sign in with your G2X account; its permissions apply to every tool. Your AI agent or model subscription is separate.

The service applies usage and request limits. Each plan also sets a pace, a number of requests a minute, shared by all of your connections. Checking your usage or an analysis run's status does not count toward it. When a request goes past a limit, it returns HTTP 429 with a message that names the limit and when to try again, for example “You've reached your plan's limit of 30 requests a minute. Try again in 2 seconds.” Wait that long; repeating the request sooner will not increase your access.

Open your G2X account menu to check Lumen usage. An agent can report the balance only when its connection provides an account-usage capability. A request limit is not a price, and it is not a measure of your remaining Lumen usage. Do not estimate an account balance from the number of tool calls. Machine-readable access information.

Errors & recovery

• 401: Reconnect using the authentication method configured in your client. If sign-in keeps failing, contact G2X support.
• 429: You reached a limit. The message says which one and when to try again; wait that long and do not retry in a loop.
• 503: G2X is busy right now. The request did not run; try again in a few seconds.
• Tool error: Inspect isError and the returned message. Check the input schema, your access and the required preceding workflow step.
• 405 on GET /mcp: Use MCP JSON-RPC POST requests with Content-Type: application/json and Accept: application/json, text/event-stream. The endpoint is not a browser page.

When no record matches, refine the question instead of inventing a result. Use bounded retries and keep run IDs for analysis requests.

Machine discovery

• Agent skill manifest: complete package and file checksums
• llms.txt: product and documentation index
• llms-full.txt: complete text guide
• MCP server manifest: remote transport
• MCP server card: domain discovery
• API catalog: interface documentation
• Sitemap: public pages

Request the home page or this guide with Accept: text/markdown for the text version. Responses include Vary: Accept.

Versioning: the MCP handshake negotiates the protocol version. Tool schemas are discoverable once you are signed in, so review them when you upgrade your client.