Your test lab in your AI agent
TestLabIQ ships its own MCP server. Connect any MCP-capable AI agent – Claude, ChatGPT, Cursor or your own – and ask your lab questions in plain language – or let it handle routine work directly.
Book a demoWhat you can do with it
Ask instead of click
“Which calibrations expire next month?” or “Which sequences are behind schedule?” – answered straight from your live data.
Delegate routine work
Have projects, test sequences, test steps, samples or milestones created and maintained – from an email, a spec or a spreadsheet.
Diagnose failures
Why didn't the measurement delivery arrive? Why did the report fail? The agent reads the stored reason – no screenshot needed.
Built secure – not bolted on
Same permission model
Every access has a role (viewer, member, admin) and only sees your organization – exactly like a human member.
More in the help centerOrganization can't be hijacked
No tool takes the organization as a parameter. Even an injected instruction can't steer the agent into other data.
More in the help centerComplete audit log
Every agent action appears in the audit log with actor type “agent” – as traceable as any human change.
More in the help centerNo deletion, no permissions
Deleting, member permissions and security-relevant settings are deliberately out of reach for agents.
More in the help center60+ tools
Read and maintain projects, sequences, test steps, stations, samples, equipment, customers, milestones, nonconformances and chain of custody.
More in the help centerBuilt-in help search
The agent searches this help documentation itself and explains features in the context of your data.
More in the help centerConnected in one minute
No vendor lock-in: Claude, ChatGPT, Cursor, VS Code or your own agent – enter the address, sign in in your browser, choose organization and permissions:
URL https://app.testlabiq.com/api/v1/agent/mcp
Transport Streamable HTTP
Sign-in OAuth in the browser – or x-agent-token headerNo token needed: the client discovers our sign-in on its own. For automations without a person in front, there are fixed agent tokens. Example configs are in the agent access help article.
Frequently asked questions
What is MCP?
The Model Context Protocol is an open standard that lets AI assistants from different vendors access external tools and data. TestLabIQ provides an MCP server for it – no programming required.
Which AI clients are supported?
Any MCP-capable client speaking Streamable HTTP – via OAuth sign-in e.g. Claude (web, desktop, Claude Code), ChatGPT connectors, Cursor and VS Code with GitHub Copilot; via a fixed token in a header also Codex CLI, scripts and custom agents built on the MCP SDKs.
Can the agent delete data?
No. Deleting, member permissions and settings like webhooks or API keys are deliberately not available as tools. With the viewer role, access is read-only.
How do I get access?
For AI assistants, yourself: enter the address in your client, sign in to TestLabIQ, choose organization and permissions – done. The app acts on your behalf and never with more rights than you; disconnect it anytime under “Connected apps”. We set up fixed tokens for automations.
Is it the same access as the REST API?
The fixed agent token: yes, it also authenticates the REST API, so scripts and AI agents share the same permissions. An OAuth connection, by contrast, only applies to the MCP server.