BrightMove now runs a live Model Context Protocol (MCP) server at mcp.brightmove.com, giving AI clients direct, structured access to real ATS data, the same objects documented in BrightMove’s Public API Reference. That is a meaningful shift from how ATS integrations have traditionally worked, and it lines up directly with what we published in our AI Done Right playbook: AI should live inside the recruiting workflow, not bolted on beside it.
What Is MCP, and Why Does It Matter for an ATS?
MCP is an open standard, introduced by Anthropic, for connecting AI systems to the tools and data they need to act on. Instead of every AI application needing a custom, one-off integration for every system it touches, MCP gives AI clients a consistent way to discover what an API can do and call it directly. For a recruiting platform, that matters because the alternative, point-to-point integrations built and maintained by hand, is exactly the kind of fragmentation our AI Done Right playbook calls out as where recruiting AI goes wrong.
BrightMove’s server, named brightmove-public-api-mcp, connects over MCP’s StreamableHTTP transport and exposes the same objects as the Public API, so any MCP-compatible AI client can query and act on live BrightMove data directly, instead of a developer hand-coding each request or an AI tool working from a stale export.
What BrightMove’s MCP Server Connects To
The live tool set behind mcp.brightmove.com spans the full arc of a candidate’s journey through your pipeline, with dedicated read, write, and search tools for each object:
- Applicants: get, update, delete, list, create, and search applicant records
- Requirements (Jobs): get, update, delete, list, and create job requirements, plus assign or remove recruiters, sourcers, vendors, workgroups, categories, functions, and industries
- Contacts: get, update, delete, list, create, and search the hiring-manager and client-side contact record
- Clients and Offices: get, update, delete, list, and create the organizational structure a job and its stakeholders sit inside
- Opportunities and Placements: get, update, delete, list, and create records across the opportunity-to-placement arc
- Submittals and Submittal Activities: get, update, delete, list, and create both the submission itself and the activity log tied to it, the record of what happened and when
- Universal Search, Lookup, and Custom Fields: search across every object type, pull reference data (workgroups, states, placement types, pay types, job categories, functions, industries), and set, clear, or get custom field values
That is not a documentation lookup tool. It is a live, callable interface into the same data your recruiters see in the ATS every day, including the Submittal Eval scoring engine referenced in our resume-tips content for job seekers, and the prompt-metadata context an AI agent needs to reason about a candidate or a role accurately instead of guessing.
AI Done Right: Tethered, Not Bolted On
Our AI Done Right framework defines four things AI has to be to actually work inside recruiting: Tethered to the workflow, Enriched with context, Action-oriented, and Measured by outcomes. A live MCP server is, in a very literal sense, the technical mechanism for the first pillar. Instead of an AI tool living in a separate tab, exporting a CSV, or getting fed a stale snapshot of your data, an MCP-connected AI client reads and acts against your live ATS data through the same schema your own team uses every day.
That also directly serves the second pillar, context. An AI agent that can query a candidate’s submittal history, a job’s requirements, and a contact’s prior activity in the same conversation is working with the same institutional memory a tenured recruiter has, not a blank slate.
The ATS as the System of Record
Every candidate conversation, every hiring manager note, every status change and submittal activity in BrightMove gets logged against the submittal, the applicant, or the job it belongs to. That is not incidental. It is what makes the ATS the single, authoritative historical record of every communication and activity between a candidate, a hiring manager, and a recruiter, not just at the moment it happens, but as a permanent, queryable history.
That system-of-record role is exactly why the live tool set behind mcp.brightmove.com matters so much. An AI agent, whether it is helping a recruiter draft a follow-up, summarizing a candidate’s status for a hiring manager, or flagging a submittal that has gone quiet, is only as good as its access to that history. MCP gives an AI client a direct, standardized path to the same submittals, submittal activities, notes, and status changes your team already relies on, instead of a second, disconnected copy of the truth living somewhere else.
What This Means for Recruiters and RPO Firms
For RPO firms and staffing agencies running high volumes of requisitions across multiple clients, this closes a gap that traditional integrations leave open. A Zapier-style integration is excellent at moving data between systems on a trigger. An MCP connection is different: it lets an AI client ask questions and take action, creating a submittal, updating a status, logging an activity, against live BrightMove data in the same conversation, using the same objects (applicants, submittals, activities, requirements) your recruiters already work in every day.
For developers and technical teams, the practical benefit is immediate: an AI coding assistant or internal tool can call BrightMove’s actual, current data and schema directly through mcp.brightmove.com, instead of stitching together REST calls by hand for every new integration.
Frequently Asked Questions
What is MCP, and why does it matter for an ATS?
MCP, the Model Context Protocol, is an open standard for connecting AI systems directly to the tools and data they act on. For an ATS, it replaces bespoke, one-off integration work with a consistent, discoverable way for AI clients to query and act on recruiting data.
Does BrightMove have an MCP server?
Yes. BrightMove runs a live MCP server, brightmove-public-api-mcp, at mcp.brightmove.com, exposing read, write, search, and list tools across applicants, requirements, contacts, clients, offices, opportunities, placements, submittals, and submittal activities, the same objects documented in BrightMove’s Public API Reference.
How is MCP different from BrightMove’s Zapier integration?
Zapier connects BrightMove to other apps through triggered, no-code automations, ideal for moving data on an event. MCP is built for AI clients specifically: it lets an AI agent query and act on live BrightMove data conversationally, using the same schema and objects your team already works in.
Is BrightMove’s ATS really the system of record for candidate communications?
Yes. Every submittal activity, note, and status change tied to a candidate, a job, or a contact is captured against that record in BrightMove, creating a complete, queryable history of communications and activity between candidates, hiring managers, and recruiters.
Do I need to build something custom to use BrightMove’s MCP server?
No. Any MCP-compatible AI client can connect to mcp.brightmove.com directly to discover and call BrightMove’s available tools, without a custom integration project.
Curious what AI done right looks like inside your own ATS? Take the AI Done Right assessment or schedule a demo to see BrightMove’s platform, and its live MCP server, in action.
Related reading: AI Done Right | BrightMove + Zapier: Automating the RPO Tech Stack | BrightMove Developer Documentation