Discera, as tools your assistant can call.
Discera runs a Model Context Protocol server, so Claude Desktop, claude.ai, and ChatGPT can read your reports, start analyses, and manage schedules directly — authorized with OAuth, scoped to your workspace, and fully audited.
An open protocol, not a plugin.
The Model Context Protocol is a standard way for AI assistants to call external tools. Discera implements it as a server, so any MCP-capable client can connect — no bespoke plugin per assistant.
Once connected, your assistant can reach a curated catalog of Discera tools: read and summarize reports, describe the filters behind them, preview and start an analysis, check or cancel a job, and run the full schedule lifecycle.
It’s the same Discera with the same guardrails. Every session is bound to your workspace and permissions, starting an analysis takes a confirmation, plan limits still apply, and anything that changes state lands in the audit log.
What your assistant can do.
- 01
Read your workspace
List and open reports, describe the filters behind them, and pull your workspace profile, prompt library, and cost history.
- 02
Run analyses
Preview an analysis to see scope and cost, then start it — with a confirmation step before anything actually runs.
- 03
Track and manage jobs
Check the status of a running analysis or cancel it, all from the assistant you're already in.
- 04
Manage schedules
List, create, activate, pause, resume, or delete recurring analyses — the full schedule lifecycle as tools.
Three steps to a live connection.
- 01
Add the server
Point your MCP client at mcp.discera.ai. Claude Desktop, claude.ai, and ChatGPT all speak the Model Context Protocol.
- 02
Authorize once
Connect with OAuth 2.1 and PKCE. Discera verifies your identity and binds the session to your workspace — no credentials stored in the client.
- 03
Ask in plain language
Your assistant now calls Discera's tools directly. Ask for a report, kick off an analysis, or set up a schedule by chatting.
Open to connect, locked to you.
- Standard MCP
An open Model Context Protocol server over streamable HTTP — any MCP-capable client can connect.
- OAuth 2.1 · PKCE
Authorize with modern OAuth. The session is bound to your workspace and your permissions.
- Confirm to run
Starting an analysis requires a single-use confirmation token, and the same plan limits always apply.
- Audited
Every mutating tool call is written to your workspace audit log, tagged as MCP activity.
Frequently asked.
What is the Discera MCP server?
It's a Model Context Protocol server that exposes Discera's reports, analyses, and schedules as tools an AI assistant can call. MCP-capable clients like Claude Desktop, claude.ai, and ChatGPT connect to it directly.
Which clients can connect?
Any client that supports the Model Context Protocol — including Claude Desktop, claude.ai, and ChatGPT. You connect by pointing the client at the Discera MCP endpoint and authorizing with OAuth.
Is it safe to let a client start analyses?
Yes. The connection is scoped to your workspace and permissions, starting an analysis requires a single-use confirmation token, your plan limits still apply, and every mutating call is recorded in the audit log.
How is this different from the AI Agent page?
The MCP server is the connection surface — the protocol and tool catalog that external assistants use. The AI Agent page covers the broader experience of driving Discera by conversation, including the built-in in-app assistant.
How does authentication work?
Through OAuth 2.1 with PKCE. You authorize from your client, Discera verifies your identity, and the session is bound to your workspace. No long-lived credentials are stored in the client.
Bring Discera into your assistant. Over an open protocol.
Point any MCP client at Discera, authorize with OAuth, and run your reports and analyses by conversation — free to try.
No credit card required for the 30-day trial · Most first reports back before the next stand-up