HubSpot Will Not Charge Extra When Claude or ChatGPT Works Its CRM
HubSpot says each paid seat now includes everyday CRM access from Claude, ChatGPT, Gemini and Copilot at no extra cost. Agent work at scale, such as enrichment and personalised outreach, is billed in credits.

HubSpot has set out how it will charge for AI agents that work inside its CRM from outside tools. In a post on its company news site, updated on 18 November 2026, the company says every paid seat now includes enough "headless" access for everyday work done from Claude or ChatGPT, at no additional cost.
Headless here means using HubSpot without opening HubSpot: a marketer asks Claude, ChatGPT, Gemini or Microsoft Copilot to look up a contact, update a deal or pull a report, and the assistant does the work through HubSpot's connectors, its MCP server or its Agent CLI. Until now, the open question for buyers was whether that kind of usage would be billed as a new product. HubSpot's answer is that it will not, for normal volumes.
The company does draw a line. Agents that do heavy work at scale inside HubSpot's own governed agent environment, such as enriching records or personalising outreach, are paid for with HubSpot Credits.
Key facts
- Included with each seat: connectors for Claude, ChatGPT, Gemini and Microsoft Copilot, plus the HubSpot MCP server and Agent CLI
- No credits for API use: third-party agents calling HubSpot's existing APIs use the API capacity already in the plan
- Paid with credits: agents working at scale in HubSpot's governed agent environment, such as enrichment and personalised outreach
- No new seat type: AI access does not require a separate or upgraded seat
- Not disclosed: exact API limits and the price of extra capacity
What is free and what costs credits
HubSpot splits agent usage into two buckets. The first is everyday work done from an outside assistant. Connecting Claude or ChatGPT to the CRM, asking questions about records and making routine updates is covered by the seat. When a third-party agent goes through HubSpot's existing public APIs, those calls count against the API capacity that each plan already includes, not against credits.
The second bucket is work done at scale inside HubSpot's governed agent environment, which the company says runs on purpose-built agentic APIs. Enriching large numbers of records or personalising outreach across a whole segment falls into this category, and it is billed in HubSpot Credits. Power users can also spend credits to raise the capacity attached to their own seat, without buying an add-on or switching seat type.
HubSpot says its plans continue to include generous API capacity and that teams can add more at any time, but the post gives no figures for either the included capacity or the cost of extra volume.
Why the pricing matters to email teams
For marketing teams, the CRM is where segments, lifecycle stages and consent data live. A growing share of the work that used to happen in the HubSpot interface, such as building a list, checking why a contact did not get an email or drafting a campaign brief from deal data, can now be done by asking an assistant. Pricing that work as part of the seat removes one reason to keep these tasks inside the vendor's own AI features.
The credit line is where costs can grow. Personalised outreach at scale is exactly the kind of job email teams want to automate, and HubSpot puts it in the metered category. Teams planning AI-written personalisation across large lists should budget credits for it, rather than assuming it is covered by the seat.
Part of a wider move to MCP
HubSpot is not alone. Over the past weeks, several email and marketing platforms have opened their data to outside assistants through MCP servers, including Iterable, ActiveCampaign, Rule and Selzy. What sets HubSpot's announcement apart is that it addresses pricing directly, which most vendors have so far left vague.
What to watch
Two details will decide how generous the model is in practice: the size of the API allowance in each plan and the price of extra capacity, neither of which is in the announcement. Governance matters too. An assistant connected through MCP acts with the permissions of the user who connected it, so the usual rules on who can export contacts or change consent fields still apply. A connected assistant that can edit subscription status can also break a suppression list if nobody reviews its changes.
What it means for senders
What changes
HubSpot customers can run everyday CRM work from Claude, ChatGPT, Gemini or Copilot at no extra cost, while large-scale agent jobs such as enrichment and personalised outreach are billed in credits.
Who is affected
HubSpot customers on paid seats, especially marketing and email teams planning to automate list building, enrichment or personalisation with AI agents.
When
Announced in a company post updated on 18 November 2026; HubSpot did not give separate effective dates.
Action checklist
- Connect your assistant through HubSpot's official connector or MCP server rather than unofficial scripts, so usage stays inside the included capacity.
- Separate everyday tasks from bulk jobs and estimate credits for any enrichment or personalised outreach at scale.
- Ask your account manager for your API allowance before planning high-volume agent workflows.
- Review user permissions on exporting contacts and editing consent properties before connecting assistants.
- Log what agents change in lifecycle stages and subscription status, so list problems can be traced.
Sources
Related on SMTPedia
SMTPedia News follows our editorial standards. Spotted an error? See how we correct it.
Get SMTPedia news first on Google
Add us as a preferred source to see our stories in Top Stories and Discover.
Add as preferred sourceUseful story?
SMTPedia news is free and independent. A coffee helps us keep covering what changes for senders.
Buy SMTPedia a coffeeNewsletter