> Connecting a destination — HubSpot, Salesforce, Apollo

TL;DR: Settings → Destinations → + Connect. Pick the destination, OAuth in, configure scope and defaults. Done — any audience can now push to it.

Expand

HubSpot

1. Click + Connect → pick HubSpot
2. OAuth flow → log in to your HubSpot account, authorize Unstuck Engine
3. Pick the HubSpot Account to push to (if your HubSpot user has access to multiple)
4. Default object — Contact (default), Company, or Deal. Most workspaces use Contact for lead-level pushes
5. Default list (optional) — every push lands in a HubSpot list with this name; you can override per audience
6. Field mapping defaults — work_email → email, first_name → firstname, etc. Edit the mapping inline
7. Test sync — Unstuck pushes a dummy record; you confirm it appeared in HubSpot

Scope requested by Unstuck Engine: contacts.read, contacts.write, lists.write, companies.read, companies.write. No deal write by default — enable separately if you push to Deals.

Salesforce

1. Click + Connect → pick Salesforce
2. OAuth flow → SFDC login, authorize
3. Pick the Org (production / sandbox)
4. Default object — Lead or Contact. Picking Lead is standard for net-new outbound; Contact for accounts already in your CRM
5. Default campaign (optional) — every push lands in this Campaign as a CampaignMember
6. Field mapping defaults — same idea, SFDC field names
7. Owner assignment — auto-assign to a queue, a specific user, or by round-robin
8. Test sync — same dummy-record verification

Scope: api access, contacts/leads read+write, campaigns read+write. Refresh token configured for long-running sync.

Apollo

1. Click + Connect → pick Apollo
2. Apollo doesn't use OAuth — paste your Apollo API key
3. Default sequence — every push enrolls members in this sequence
4. Mailbox routing — if your Apollo workspace has multiple sending mailboxes, pick the round-robin policy
5. Field mapping defaults — usually minimal (email, first_name, last_name, company, title, linkedin_url)
6. Test sync — adds a dummy record to a test sequence you verify

Apollo's API rate limits apply — Unstuck Engine respects them and batches sync at rate-limit-friendly intervals.

Smartlead

Same flow as Apollo, with Smartlead's API key. Default sequence is mandatory (a sync target).

Custom webhooks

For anything else (your own internal tool, a custom AI agent, a notification system):

1. Click + Connect → pick Custom webhook
2. Enter your webhook URL
3. Set an auth header (e.g. Authorization: Bearer <your-secret>)
4. Pick the events to POST — member_added, member_removed, member_updated, all of the above
5. Test with a dummy event

Your webhook receives a JSON payload per event. Schema is documented in the Edge Function Contracts spec.

Verifying the connection

Each connected destination has a status indicator on the Destinations page:

• 🟢 Connected & syncing
• 🟡 Connected, no audiences using it
• 🔴 Connection error (OAuth token expired, rate-limited, schema mismatch)

Click into any destination to see the last 10 sync events, success / failure counts, and last error message.

Disconnecting

Settings → Destinations → row → Disconnect. Stops all syncs and revokes the OAuth token. Records already in the destination remain; Unstuck Engine doesn't clean up downstream.

Related

• Field mapping
• Pushing an audience to a destination
• Using the Signals API
• Using the Enrichment API