Workflow tools
Use these first when your request maps to a GTM job rather than one API call:Target account people search
find_people_at_target_accounts accepts company domains, company names, or company LinkedIn URLs. Prefer domains or names when you know the account. A LinkedIn URL is useful when you already know the exact company page, but the tool should not silently replace the requested account with a different company page.
When the workflow waits for the people-search operation to finish, the underlying result can include:
If the workflow returns an operation ID instead of final people, call
get_operation with that ID and read the same fields from the operation result.
API tools
Use these when you want direct control over a specific public API endpoint:Example prompts
Async results
Many The Hog API calls run asynchronously. MCP tools may return an operation ID, poll URL, child operation IDs, warnings, and partial step results. If a tool returns an operation ID instead of final data, ask your client to callget_operation with that ID.
Deep-research-bearing tools (research_company, research_person, scrape_and_extract, start_deep_research) are async-first: because deep research runs for minutes, they return an operation ID immediately (status queued) rather than blocking. Poll get_operation until the status is completed to fetch the result. You may set waitForResult: true to wait inline, but the wait is capped at ~50s (below the ~60s MCP client/gateway ceiling); if it does not finish in time you still receive the operation ID to poll. Re-attaching with get_operation does not consume additional credits — do not re-issue the call to retry.
Slow LinkedIn profile post scrapes use the same pattern: list_linkedin_profile_posts queues work and returns an operation ID immediately. Poll get_operation until status is succeeded to fetch the posts.