Multi-step agents
Run autonomous workflows on Remote data. The agent runs, the work gets done.
Everything you manage in Remote is now accessible to any AI tool your team uses. Instead of exporting spreadsheets, building reports, or stitching together VLOOKUPs, HR and Payroll teams can now query live workforce data directly from chat.
MCP turns Remote into a live data layer any AI agent or app can build on. Build dashboards, audits, multi-step workflows, and entirely new tools — all running on real workforce data.
Run autonomous workflows on Remote data. The agent runs, the work gets done.
Combine Remote data with Salesforce, Notion, Google Drive, or anything else your AI tool can reach.
You sign in. That’s it. The authentication, the permissions, and the compliance posture all carry over.
Sign in through your browser, the same way you log into Remote. We don't ask you to generate API keys or store secrets, and your credentials never touch the AI tool you're using.
Whatever permissions you have in Remote, you have through MCP. An employer admin sees employer data; an employee sees their own, and nothing more.
MCP exposes specific, structured endpoints, not raw database access. Remote controls what's returned, so AI tools can't pull bulk exports or unstructured data dumps.
Remote sends only what you ask for, and only to the AI tool you've signed into. Your workforce data isn't used to train any AI provider's models.
MCP runs on Remote’s existing security and compliance posture, with the same rigor you already trust.
Open your client, paste a URL, sign in. That’s it. Your assistant now has access to whatever Remote tools your role allows. No API tokens, no JSON config, no admin tickets.
These clients work today or ship soon—desktop apps, IDEs, the CLI, or anything that speaks MCP.
Paste one URL, into your tool’s connector or MCP config. Walkthroughs for each tool below.
Two quick logins: Claude, then Remote. Your role determines which tools become available. After that, just type.
The browser app — quickest path, no install.
The macOS / Windows app. Connectors sync with claude.ai — set it up once, both clients get it.
For the terminal-native crowd. One command, OAuth handled in your browser.
claude, then ask “list my last three payslips”.claude mcp list — you should see remote with status connected.
HTTP MCP via a config file. Same shape works in Continue, Zed, Windsurf, and most other IDEs.
.cursor/mcp.json in your project root (or ~/.cursor/mcp.json globally) with:~/.continue/config.json), Zed (settings.json), and Windsurf — paste it into their respective MCP block.
For any client that speaks the Model Context Protocol — current or future.
https://mcp.remote.com/mcp2025-06-18 or current spec.If your product runs on Remote (through embedded employment, integrations, or custom APIs) MCP gives your customers a new AI-native way to interact with the workforce data your product manages. Your existing API integrations keep doing what they do. MCP layers on top.
Workday, Gusto, BambooHR, Personio, and others already provide global employment for their customers through Remote. MCP extends that to AI tools, without anything to rebuild on your end.
Continue using Remote's public API for the durable read/write integrations you've built. MCP is the AI-native complement.
Remote MCP uses OAuth 2.0 with PKCE and dynamic client registration. There's nothing to provision, store, or rotate. Your client registers itself, the user signs in through their browser, and access is scoped to their existing Remote role automatically.
Sign in, follow the quickstart, and start working in real time.