Guides · carbon.talent
Hire data talent from Claude
Last updated 18 Oct 2026
You can post a data role and review applicants from Claude by connecting it to carbon.talent through Carbon AI. Carbon AI is a Model Context Protocol (MCP) server at https://mcp.carbondatasolutions.com/mcp, available on Team and Scale.
To set it up, add the address as a custom connector in Claude, then approve the connection in carbon.talent. After that you can ask things like "draft a senior data engineer role from this job ad" or "who applied to my analytics engineer role this week?".
What you can do from chat
- Create a role draft. Give a title and details, paste a job ad or pass a link. Only what the ad states is filled in.
- Post a role. Most roles go live right away; some are held for a quick review.
- List your roles or get one role. You see status, location, engagement, skills, salary range and applicants by stage.
- List applicants or get an applicant summary. You see:
- stage history;
- skills and years of experience;
- location;
- real-world data experience test status;
- interview status;
- your team's notes.
- Search the talent catalog by skills, experience, location or keywords.
- Move an applicant to Applied, Screened, Interview or Offer.
What it doesn't do (by design)
- It doesn't return email addresses, phone numbers, CV files, profile links, rates or scores. Those stay in the portal.
- It doesn't mark anyone Hired. You do that in the portal.
- It only sees your company's roles and applicants.
How to connect
In Claude, the connector is named carbon.talent.
- Copy the server address:
https://mcp.carbondatasolutions.com/mcp. - Add it as a custom connector in Claude. Carbon AI also works with other AI assistants and tools that support custom MCP connectors, such as Cursor, VS Code and LM Studio.
- Sign in to carbon.talent and approve the connection. You can revoke it any time under Settings, AI connections.
Full details: Carbon AI.
Prompts that work well
- "Draft a contract dbt developer role, 6 months, remote, 4 hours overlap with London, from this job ad: [paste]."
- "Summarise the applicants on my Data Engineer role who are at Screened and have Airflow."
- "Search the talent catalog for senior data scientists with experimentation experience in Europe."
- "Move Priya K. to Interview."
Using AI to write the job description
Assistants write good first drafts. To get one:
- Give the assistant the outcomes, stack, engagement, overlap hours and pay. You can also start from our data engineer job description template.
- Edit the draft so it's honest about the current state of your data.
- Remove anything generic before you post.
Good practice
- A person makes every stage move and offer decision.
- Don't paste candidates' personal details into other tools.
- Use the portal for contact details and contracts.
Not on Team or Scale?
Post your role free in the portal and use everything else on carbon.talent. Add Carbon AI when you move to Team or Scale. Team plan: 30 days free, no card needed.
Questions companies ask
- Can I post a job from Claude?
- Yes, on carbon.talent's Team and Scale plans. Add Carbon AI as a custom connector in Claude and approve it in carbon.talent. Then ask Claude to draft and post your role.
- What is a recruiting MCP server?
- A Model Context Protocol (MCP) server lets an AI assistant read and act on a hiring system, with your permission. Carbon AI is carbon.talent's MCP server for data and AI roles.
- Is my candidate data safe when I connect an AI assistant?
- The assistant only sees your company's roles and applicants. Contact details, CVs, rates and scores stay in the portal. You approve the connection and can revoke it any time.
- What can't the assistant do?
- It can't mark someone Hired, return contact details or CVs, or see other companies' data.
- Which plans include Carbon AI?
- Team and Scale.