How to connect HubSpot to OpenClaw

The ClawHub skill teaches OpenClaw about HubSpot, but the calls run through the ClawLink plugin and your connected account. Make sure HubSpot is connected in the dashboard, then start a fresh chat so OpenClaw reloads the tool catalog. If OpenClaw runs as a persistent gateway, restart it so the new tools register.

Reconnect HubSpot from the dashboard, then start a fresh chat if the runtime still has the old tool catalog loaded.

HubSpot tool schemas load on demand the first time a tool runs and are cached after that, so this error usually clears on its own: wait a few seconds and retry the same request. If every HubSpot call keeps failing with it in a fresh chat, reconnect from the dashboard, and contact support if it still persists — that pattern points to a configuration problem on our side, not something you can fix by reconnecting again.

Two usual causes. The connected account may not have access to the specific workspace, inbox, store, or project in the request — check that first. If access looks right, the agent may have sent a placeholder value (like "YOUR_ID" or an example id from documentation) instead of a real one: ask it to run a list or search tool first, then retry the action with a real id from those results. Most failures at this stage are one of these two, not ClawLink bugs.

This is the case the search results answer worst and the one this page is asked about most. Work through it in order. First, user permissions: HubSpot's model is granular, so the connected user may lack access to that object type, that pipeline, or records owned by other people, and the failure looks like a broken integration while everything else works. Second, the scope of the grant — a partial approval leaves a connection that reads but cannot write, so if every create fails and every search succeeds, reconnect and accept the full set. Third, the portal, which is the sandbox-versus-production trap. Last, a placeholder: HubSpot object ids are numeric, and an id lifted from documentation fails as a permission error rather than a not-found, so have the agent search first and use a real id.

Ask the agent to diagnose it:

Call hubspot_get_account_info and hubspot_retrieve_owners, and tell me which portal this connection is on and which user it acts as. Then run hubspot_search_contacts_by_criteria for anything at all and tell me if results come back. Do not retry the failed call yet.

Creating is a single call and searching is not, so an agent told to "add this person to HubSpot" will often do exactly that without checking. HubSpot's own deduplication keys on email for contacts and domain for companies, which means a record created with a slightly different email address becomes a genuine duplicate rather than being merged. The prompt-level fix is to make the search explicit: ask for hubspot_search_contacts_by_criteria on the email first, then create only if nothing comes back. If duplicates already exist, merging is a HubSpot-side operation rather than something to ask the agent to improvise.

Ask the agent to diagnose it:

Search HubSpot for this email with hubspot_search_contacts_by_criteria and show me every match, including the record ids and created dates, before creating anything.

Two causes. The first is the criteria shape: HubSpot's search filters on properties with specific operators, so a search that names a property the portal does not have, or uses a value in the wrong format, returns an empty set rather than an error. The second is record ownership and visibility — a user limited to their own records genuinely cannot see everyone's, so the agent's empty result is accurate for that user. Ask it to search on email or domain first, which are the most reliable keys, before assuming anything is wrong with the connection.

Try reconnecting HubSpot and complete the consent flow in the same browser session. Partial OAuth approvals or switching accounts mid-flow can leave the connection incomplete.

Yes. ClawLink is the fastest way to connect OpenClaw to HubSpot: link your HubSpot account once in the browser and OpenClaw can call the HubSpot API through 224 ready-made tools — no custom code or token handling.

Paste the setup prompt from this page into OpenClaw. It installs the ClawLink HubSpot skill from ClawHub, then you click Connect in the dashboard to authorize HubSpot. OpenClaw calls the tools from the next message — no config files or API keys to manage.

About two minutes. Sign in, click Connect next to HubSpot in the dashboard, authenticate, and OpenClaw can use it from the next chat message.

The alternative to ClawLink is usually manual OAuth app setup plus your own token handling, permission troubleshooting, and tool plumbing for OpenClaw. That is fine if you want to build and maintain the integration yourself. Most teams just want HubSpot working from chat.

Neither. HubSpot deprecated API keys, and the usual replacement is creating a private app in your HubSpot account, choosing scopes, and copying its access token into whatever is calling the API. ClawLink uses HubSpot's OAuth flow instead, so you approve the connection on HubSpot's own screen and the credential stays on the ClawLink server. The practical difference is what happens when things change: a private app token sits wherever you put it until someone rotates it, while an OAuth grant can be revoked from the ClawLink dashboard or from HubSpot's connected-apps settings and stops working immediately.

The portal you selected during the connect flow, with your user's permissions inside it. That matters more in HubSpot than in most tools because permissions there are genuinely granular — a user restricted to their own contacts, or without access to a particular pipeline, produces an agent with exactly those limits. If your organisation has several portals, for example a sandbox and production, check which one you approved before concluding that records are missing; a connection pointed at the sandbox looks completely healthy and finds nothing you recognise.

The connection is HubSpot's own OAuth flow, so the permissions are approved on HubSpot's screen, no password is handled, and access can be revoked from either side at any time. The risks worth planning for are about your CRM rather than the mechanism. The agent can create and update records, so a vague instruction can produce duplicate contacts or overwrite a field somebody curated — ask it to search before it creates, and to show you what it will change. And CRM notes contain text written by other people, so treat what it reads as untrusted input if you are going to let it act on that basis.

Considerably more, and this is where the toolset differs from the small CRM connectors. Alongside contacts, companies, deals, tickets, tasks and owners, it reaches marketing email, A/B variations and aggregated statistics, imports and their error details, pipeline change audits, and generic CRM object search that covers custom objects. The practical consequence is that reporting questions and data-hygiene work are in scope, not just record creation — which is usually where an agent earns its place in a CRM.

The ClawHub skill teaches OpenClaw about HubSpot, but the calls run through the ClawLink plugin and your connected account. Make sure HubSpot is connected in the dashboard, then start a fresh chat so OpenClaw reloads the tool catalog. If OpenClaw runs as a persistent gateway, restart it so the new tools register.