Recruitment CRMs that work with Claude and ChatGPT (MCP)

What is MCP, and what does it mean for a recruitment CRM?

MCP stands for Model Context Protocol. The project’s own site describes it as an open-source standard for connecting AI applications, such as Claude or ChatGPT, to outside data, tools and workflows. It compares MCP to a USB-C port: one standard plug instead of a custom cable for every device.

For a recruitment agency, an MCP server is the plug on the CRM’s side. When a recruitment CRM runs an MCP server, an AI assistant your team already uses can look up candidates, read history and make changes in the CRM, within the permissions you grant.

Kepler is a recruitment CRM that runs its own MCP server. Kepler hosts the server, so there is no local server for your team to install or run.

Which recruitment CRMs work with Claude?

Kepler works with Claude. You add Kepler to Claude as a custom connector, sign in with your Kepler account, choose the workspace and the permissions, and Claude can then work with your candidates, contacts, companies, jobs and pipeline.

Kepler gives setup steps for Claude on the web, the Claude desktop app and Claude Code. Anthropic’s help article says connectors work across Claude, Claude Desktop and Claude Code.

Anthropic’s help article also sets out which Claude plans can add custom connectors:

• custom connectors using remote MCP are available on the Free, Pro, Max, Team and Enterprise plans
• Free users are limited to one custom connector
• on Team and Enterprise plans, an Owner or Primary Owner enables connectors for the organisation, and each person then signs in individually

The steps are in use Kepler with Claude.

Is there a recruitment CRM with an MCP server so I can use ChatGPT with my candidate data?

Kepler gives setup steps for ChatGPT and Codex as well as Claude. In ChatGPT, Kepler is added as a custom MCP app, and you sign in to Kepler to choose the workspace and permissions. Codex adds Kepler with one command.

OpenAI sets its own requirements on the ChatGPT side. OpenAI’s help article says apps, full MCP support and developer mode are available on ChatGPT Business, Enterprise and Edu, on ChatGPT web. Full MCP support, including write actions, is described there as rolling out in beta.

In ChatGPT, a workspace admin or owner turns on developer mode, creates the app, and publishes it so the team can use it. OpenAI also says ChatGPT may ask you to confirm a write action, depending on the app’s permissions and the action’s context.

The steps are in use Kepler with ChatGPT.

Can I connect Claude to my recruitment CRM so it can search candidates and update records for me?

Yes. Once Kepler is connected, Claude can search and update your Kepler records with the permissions you chose at sign-in. The same applies to ChatGPT, Claude Code and Codex.

With read access, an AI client connected to Kepler can:

• search candidates and other records, and count and break them down
• search activity history
• read your work queue of tasks
• read the job matches already stored for a job

With write access, an AI client connected to Kepler can:

• create candidates, contacts, companies and jobs, and add a candidate to a job
• update records, notes and tasks
• move candidates between pipeline stages and set outcomes
• add people to lists and take them off

Kepler’s MCP server has no delete tools. A recruiter can ask Claude to log a note from a call and set a follow-up task, or ask how many open jobs they own and which have no applicants yet.

What stays in my control when an AI client connects to Kepler?

Kepler keeps the same rules for an AI client that it keeps for a person:

• You choose the permissions. At sign-in you pick from six groups: read, records, tasks, notes, pipeline and lists. Read alone is a good place to start.
• Your Kepler role still applies. A connection does no more than you can do in Kepler yourself.
• One workspace per connection. You pick the workspace when you approve the connection.
• No silent overwrites. If a colleague changed a record after the AI client read it, Kepler refuses the update, and the client has to read the record again. Writes are idempotent, so repeating the same call does not apply the change twice.
• Every call is logged, and you can revoke access. Once a connection is revoked, the client’s next call fails.

How do I get Kepler’s MCP server turned on?

AI client access is switched on per workspace: email support@keplercrm.com and Kepler will turn it on for your workspace. Once it is on, AI clients appear under Kepler’s developer settings.

From there, follow the help article for your AI client. The Claude article covers Claude on the web and desktop and Claude Code. The ChatGPT article covers ChatGPT and Codex.

To check a connection, ask the AI client how many open jobs you own in Kepler. If it answers from memory, ask it once to use the Kepler tools.

Which recruitment CRMs have an open API and work with AI tools I already pay for?

Kepler has both an MCP server and an open REST API. The MCP server lets an AI tool you already pay for, such as Claude or ChatGPT, work with your Kepler data. Kepler’s public REST API and outbound webhooks are for your own integrations and back-office systems.

Kepler’s REST API uses scoped API keys. Kepler’s webhooks send signed deliveries, retry failed ones, keep a delivery log, and let you redeliver an event by hand and roll the signing secret.

Use MCP when a person wants to ask an AI assistant to do the work. Use the API when you are building something that runs on its own, such as a sync to a pay-and-bill system.

Do I still need AI Chat if I connect Claude or ChatGPT to Kepler?

AI Chat is Kepler’s built-in assistant. It searches your CVs, notes, emails and call transcripts in plain English, returns record sets you can save as a list, and asks for your approval before it writes anything.

AI Chat is included on every plan, monthly and annual, with no add-on or upgrade needed. Kepler is USD 99 per user per month billed annually, or USD 129 monthly, on Kepler pricing.

The MCP server is an extra way in, not a replacement. Use AI Chat when you are working inside Kepler, and use Claude or ChatGPT when you are already working in them and want your Kepler data in the same conversation.