Lead Generation With ChatGPT: MCP Connector Setup Guide (2026)

Lead generation with ChatGPTis possible — but only with a live data connection. Ask ChatGPT to "find me 25 VP Sales contacts at Series B SaaS companies in the US" without one and it writes back plausible-sounding names from training data. The model generates text from patterns. It has no live B2B database to query.

Connect the SyncGTM MCP and the same prompt runs a real search. ChatGPT calls find_companies, pipes the domains into find_people, pulls work emails with find_work_email, verifies them with verify_email, and writes the output to a CSV or pushes it to your CRM — all without leaving the chat.

This guide covers the complete setup for ChatGPT specifically: enabling Developer mode, creating the custom connector, authorizing the right organization, enabling the connector in each chat, and running the full lead list workflow with sample prompts and credit costs.

Why Use ChatGPT for Lead Generation?

Most AI lead generation tools wrap a database in a proprietary UI. What makes using ChatGPT with an MCP connector different is that your natural-language prompts drive structured database calls — the same reasoning ChatGPT applies to everything else now applies to your prospecting filters.

Three practical advantages over a dedicated prospecting tool:

• No-export chaining. Company domains from find_companies feed straight into find_people inside the same conversation. No CSV in between.
• Conversational filter refinement."Drop anything under 100 employees and add a minimum of $5M in funding" is one sentence, not a rebuilt filter set.
• Output in any format. The same search can produce a markdown table, a downloadable CSV, or a CRM payload — without exporting to a spreadsheet first.

The same workflow also runs inside Claude via the SyncGTM MCP. The setup path differs, but the tools, data, and credit costs are identical. See the ChatGPT vs Claude section below for a direct comparison.

Which ChatGPT Plans Support Custom Connectors?

Custom connectors — the feature that lets you add an external MCP server URL — are gated behind paid ChatGPT plans. As of 2026, the breakdown is:

If you do not see Apps & Connectors in your ChatGPT Settings, your current plan does not include custom connectors. Upgrade to Plus as the minimum paid tier to unlock this feature.

The SyncGTM MCP Tool Set

The SyncGTM MCP server exposes more than 40 tools. Four handle the majority of a standard lead generation workflow:

Search tools are billed per run, not per row. A find_companies run returning 10 results costs the same 3 credits as one returning 100. The variable cost is on the enrichment side — every find_work_email and verify_email call is billed per person. Narrow your list before you enrich it.

The full server also covers company-level data enrichment, email finding across multiple providers, LinkedIn engagement signals, job opening alerts, and tech-stack detection — useful add-ons once the core workflow is clean.

Step 1 — Enable Developer Mode in ChatGPT

Custom connectors live behind a Developer mode toggle in ChatGPT. This is a one-time setting on your account — you do not need to repeat it.

1. Sign in to chatgpt.com with a Plus, Pro, Team, or Enterprise account.
2. Click your profile icon (top-right) and go to Settings.
3. Open Apps & Connectors in the left sidebar.
4. Click Advanced settings at the bottom of the page.
5. Toggle on Enable developer features. A confirmation prompt may appear — accept it.

After enabling, you will see a new Custom connectorssection in Apps & Connectors. That is where the SyncGTM connector goes in the next step.

Step 2 — Create the SyncGTM Custom Connector

The SyncGTM MCP is a remote HTTP server authenticated via OAuth. No API key to paste, no environment file to manage, no token to rotate.

1. In Settings → Apps & Connectors, click Add connector.
2. Select Custom connector (not a pre-built integration).
3. Fill in the connector form:
   • Name: SyncGTM
   • URL: https://api.syncgtm.com/mcp
   • Authentication: OAuth
   • API key: Leave blank — SyncGTM uses OAuth only
4. Click Save. ChatGPT validates the URL and saves the connector configuration.

Connector settings:
  Name:           SyncGTM
  URL:            https://api.syncgtm.com/mcp
  Authentication: OAuth
  API key:        (leave empty)

After saving, ChatGPT shows the connector as "Not authorized" — that is expected. Authorization happens in Step 3 when you complete the OAuth flow.

Step 3 — Authorize the Right Organization

Click Authorizenext to the SyncGTM connector in Apps & Connectors. ChatGPT opens a browser tab to the SyncGTM OAuth sign-in page.

1. Sign in to your SyncGTM account in the browser tab that opens.
2. On the approve screen, review which organization is selected. If your account belongs to multiple SyncGTM workspaces, a dropdown lets you choose which one this connector draws credits from.
3. Confirm the correct workspace and click Approve.
4. Return to ChatGPT. The SyncGTM connector now shows as Connectedin Apps & Connectors.

The connector is now set up at the account level. You will not need to re-authorize unless you revoke it from SyncGTM's Connected Apps settings.

Step 4 — Enable the Connector in a Chat

This is the biggest difference between ChatGPT and Claude. In Claude, a connected MCP is available in every conversation automatically. In ChatGPT, connectors are off by default in each new chat — you enable them per conversation.

1. Start a new ChatGPT conversation.
2. Look for the connector or apps icon in the chat input toolbar (it may appear as a plug icon or a grid icon depending on your ChatGPT version).
3. Click the icon to open the connectors panel. Find SyncGTM in the list.
4. Toggle SyncGTM on. The icon in the toolbar will update to show the connector is active.

Once enabled, the connector stays active for the entire conversation. You do not need to toggle it again mid-session.

Step 5 — Check Your Credits Before Anything Else

check_credits is free and takes one second. It also confirms the connector is working — if ChatGPT returns a number, the tools are live. If it says it cannot access SyncGTM or returns a memory answer (no tool-use block), go back and verify the connector is enabled in this chat.

# Prompt:
"Check my SyncGTM credit balance."

# ChatGPT calls check_credits (shows tool-use block) and responds:
# Your SyncGTM balance is 142 credits.

Before any large run, estimate your spend: a 25-lead end-to-end workflow costs roughly 38–39 credits (3 + 3 + 25 + 7.5). If your balance is low, top it up on the SyncGTM pricing page before starting, or cap the contact list at fewer people.

Step 6 — Build and Enrich a Lead List

With the connector enabled and credits confirmed, you can run the full three-stage workflow in a single conversation. Each stage chains into the next — no intermediate exports.

Stage A — Find target companies

"Use find_companies to find 25 B2B SaaS companies in the US or Canada
with 50–500 employees that raised a Series A or Series B in the last
24 months and run HubSpot or Salesforce. Exclude staffing, recruiting,
and consulting. Return a table with: domain, company name, headcount,
last funding round, HQ country."

Credit cost:3 credits for the run. Same cost whether you get 5 results or 25. Refine the list in the same conversation ("drop anything under 100 employees") — filter changes are free, only a new search run costs credits.

Stage B — Find decision-makers

"Use find_people to find VP-level or above contacts in Sales or
Revenue Operations at the 25 companies from the previous search.
Based in the US. Limit to 25 results. Return: name, title,
company, domain, LinkedIn URL."

ChatGPT passes the domains from the previous find_companies result directly into find_people — no copy-pasting required. Credit cost: 3 credits for the run.

Stage C — Get verified work emails

Two tools run in sequence here. find_work_email uses waterfall enrichment across multiple providers to maximize hit rate. verify_email checks each result for deliverability before it lands in your sequence.

"For each person on that list, use find_work_email with their
LinkedIn URL. Then run verify_email on each email you find.
Show the results as a table: name, company, email, verify_status.
Only include rows where verify_status is 'valid' or 'risky'."

ChatGPT runs find_work_email once per person (1 credit each) and verify_email once per email found (0.3 credits each). Individual tool-use blocks appear in the UI for each call.

For a deeper look at where find_work_email fits in the broader enrichment stack, see the B2B lead enrichment guide.

Step 7 — Export to CSV or CRM

At this point you have a verified lead list in ChatGPT's working context. Get it somewhere actionable.

Export as CSV

"Export the verified leads to a CSV file with columns:
first_name, last_name, title, company, domain, email,
email_status, linkedin_url. Only include rows where
email_status is valid or risky."

ChatGPT generates the CSV as a downloadable file attachment in the conversation. Click the file link to download it directly to your machine.

Push directly to a CRM

If you have a CRM connector also enabled in this chat (HubSpot, Salesforce, or another supported integration), you can skip the CSV entirely:

"Create these as HubSpot contacts using the HubSpot connector.
Use email as the deduplication key. Set lifecycle stage
to Lead and source to Outbound — MCP."

ChatGPT calls the CRM connector's create-contact tool for each row, deduplicates on email, and reports the count of created versus existing records. No import wizard, no field mapping screen. You can also check out the best ChatGPT sales plugins for B2B teams if you want to pair SyncGTM with other connectors in the same workflow.

ChatGPT vs Claude: Key Differences for Lead Generation

The SyncGTM MCP tools, data coverage, and credit costs are identical on both platforms. The differences are in setup and connector behavior.

If you use both platforms, the MCP setup is worth doing on both. Your SyncGTM credits are shared — a search run in ChatGPT draws from the same balance as a run in Claude.

For teams that run automated or scripted lead generation workflows, Claude has an edge via Claude Code. See the B2B database MCP setup guide for the full Claude Code walkthrough.

When ChatGPT Answers From Memory (No Credit Charge)

This is the most common issue when starting out. ChatGPT is a language model first — it will answer a question from training data if it can, and only call a tool if it understands live data is needed.

How to detect a memory answer

Every real tool call produces a visible tool-use block in the ChatGPT UI before the result — a collapsible section labeled with the tool name and the parameters passed. If you receive a nicely formatted table of company names with no tool-use block above it, ChatGPT wrote that table from training data. Credits charged will be zero, and the data should not be used for outreach.

The fix — three prompts that force tool use

# Pattern 1 — Name the tool explicitly:
"Use find_companies to search for..."

# Pattern 2 — Block memory explicitly:
"Use find_companies — do not answer from your training data."

# Pattern 3 — Ask for a credit confirmation:
"After running find_companies, tell me how many credits
were charged so I can confirm it was a live call."

Pattern 1 is usually enough. If ChatGPT still answers from memory (rare, but happens when the connector is not enabled in the current chat), check that the SyncGTM connector icon is toggled on for this conversation — that is the most common cause.

Credit Cost Summary

Full cost to build 25 verified leads from scratch using the workflow above:

Search is billed per run, not per row. The 25-result cap in this guide is a workflow choice, not a cost optimization. Enrichment is billed per record — that is where narrowing the list saves money. Check current credit allowances per plan on the SyncGTM pricing page.

Conclusion

Lead generation with ChatGPT works the same way it does in Claude — connect the SyncGTM MCP, describe your ICP, and chain the tool calls from account discovery through email verification. The workflow, tools, and credit costs are identical across both platforms.

The setup difference is minor: ChatGPT requires Developer mode to be enabled once, and then the connector is toggled on per chat rather than persisting account-wide. One extra click per session in exchange for a cleaner context boundary on general-purpose conversations.

The only thing to watch on either platform is whether the model actually called the tool. Look for the tool-use block. No block means no real data — add "Use find_companies — not your training data" to your prompt and run it again. Create a free SyncGTM account to connect the MCP and run your first search today.