Comeet
HRConnect AI agents to Comeet (now Spark Hire Recruit), a collaborative recruiting applicant tracking system. Agents create positions, add candidates, and find duplicate profiles so the recruiting pipeline stays clean and current.
What This Integration Enables
Comeet, now Spark Hire Recruit, is built for teams that treat the applicant tracking system as the authoritative pipeline rather than a place resumes go to sit. Its API reflects that. Agents list positions with their status, department, employment type, and location, and create new openings against a named lead recruiter and hiring manager. On the candidate side they list candidates with their current and completed pipeline steps, retrieve a single candidate with optional extra field groups, and create candidates directly onto a position.
Two operations separate this connector from a generic ATS read surface, and both exist to keep a synced copy honest. Find Duplicate Candidates searches candidates and leads by email, phone, LinkedIn URL, or name before anything is written, so an agent can check for a match instead of creating a second profile for a person who is already in the pipeline. List Deleted Candidates returns the uids of candidates and leads that have been removed, so a system holding a mirror of that data knows which records to reconcile. Most recruiting APIs will tell an agent what exists. This one also tells it what stopped existing.
Without FlowRunner
With FlowRunner
Use Case Scenarios
Inbound Candidate Capture Without Duplicates
A referral form, a conference scanner, and an agency email all feed candidates toward the same pipeline. Instead of creating a record per source, the agent runs Find Duplicate Candidates with whatever identifiers it has, an email from the form, a LinkedIn URL from the referral, a first and last name from the scanner. When the search comes back empty it runs Create Candidate against the target position with the contact details and an internal note naming the source. When the search returns matches it stops writing and posts the candidate alongside the existing records to the recruiter in Slack. The pipeline stays one profile per person, and the duplicate check happens before the write rather than as a cleanup project later.
Requisition Publishing From an Approved Headcount Plan
Headcount approvals land in a finance system. When a role is approved, the agent reads the approval, runs List Positions to confirm no equivalent opening already exists, and runs Create Position with the role name, the status, the discreet flag, and the lead recruiter and hiring manager email addresses. Both people must already be Recruit users, so the agent verifies that first and escalates when a named hiring manager has no account rather than failing the create. The new position uid is written back to the headcount record and posted to the hiring channel, so the requisition and the approval that authorized it stay linked.
Reconciling a Warehouse Mirror With the Deletion Feed
A recruiting analytics warehouse holds a mirror of the Comeet pipeline for reporting. Each night the agent runs List Candidates filtered to the period to pick up new and changed records, then runs List Deleted Candidates over the same window. Ordinary updates are applied straight through. Deletions are different: they mean a record in the warehouse, in the sourcing spreadsheet, and possibly in an email tool should no longer be there. The agent assembles the uid list, maps each one to the systems holding a copy, and routes it for confirmation before removing anything. The mirror converges on the source of truth, and the destructive half of that convergence has a person behind it.
Human-in-Loop Highlight
Comeet is the only connector in this category that hands an agent a list of records that were deliberately deleted, and that makes the downstream purge the moment worth putting a human-in-the-loop on. Nothing about a uid in the List Deleted Candidates response tells the agent whether the deletion was a data retention request, a mistaken removal a recruiter is about to undo, or a merge. The agent has the uids and nothing else. So it does not purge. It assembles the list, resolves each uid to the mirrored records it can see, the warehouse row, the sourcing sheet line, the address in the outbound tool, and routes to the recruiting ops owner: "Comeet reports 14 candidate records deleted since the last sync. Purging them here removes 14 warehouse rows, 9 sheet lines, and 6 email addresses, and none of it comes back. Confirm the purge, exclude specific uids, or hold." Reads, pipeline updates, and duplicate checks run without asking. Destroying a record because another system said it was gone is a decision a person signs.
Agent Capabilities
7 actionsPositions
2- List Positions Lists positions (job openings) in the Comeet Recruit account, including name, status, department, employment type, and location. Supports filtering by status and by last updated time, and can include or exclude published, discreet, template, and employee only positions. Used to check for an equivalent opening before creating a new one.
- Create Position Creates a new position (job opening) in Comeet Recruit. Requires a name, a status, whether the position is discreet, and the lead recruiter and hiring manager email addresses, both of which must be existing Recruit users. Returns the created position including its generated uid.
Candidates
4- List Candidates Lists candidates in the Comeet Recruit account, including their position, current and completed pipeline steps, address, and status. Supports filtering by status and by creation time, and can optionally include deleted candidates. Used as the incremental read in a sync flow.
- Retrieve Candidate Retrieves a single candidate by their unique identifier, including position, pipeline steps, contact details, and status. Optionally requests additional field groups. Used when a workflow needs full context on one person rather than a page of summaries.
- Create Candidate Creates a new candidate on a specific position. Requires the target position uid plus first and last name, and optionally accepts email, phone, LinkedIn URL, prospect status, and an internal note. Used to capture inbound applicants and sourced prospects from forms, referrals, and events.
- Find Duplicate Candidates Searches existing candidates and leads for possible duplicates before a new record is created. Requires at least one of email, phone number, LinkedIn URL, or a first and last name combination. Returns matching records, or an empty array when there are none. Run this before Create Candidate in any inbound capture flow.
Data Retention
1- List Deleted Candidates Returns the uids of candidates and leads that have been deleted in Comeet Recruit, so they can be removed from a synced system. Supports filtering by deletion time window. This is the feed that lets a mirrored dataset stay aligned with the source, and the downstream removal it drives is routed for sign-off.
Frequently Asked Questions
What can FlowRunner do with Comeet?
FlowRunner agents can run List Positions, Create Position, and List Candidates in Comeet, plus 4 more actions.
Does connecting Comeet to FlowRunner require OAuth?
Comeet uses a custom authentication method to connect to FlowRunner.
Can Comeet trigger a FlowRunner workflow automatically?
Comeet doesn't currently expose triggers in FlowRunner. It connects as an action step inside workflows started by another trigger.
Start building with Comeet
$100 in credits. No card required. Connect in minutes.