No engineering, no API keys — sign in with the account you already have.
In your AI assistant, add a custom MCP server pointing to mcp.quanfire.ai/mcp.
Log in with your existing Quanfire account and approve access on a consent screen. Your role carries over.
Ask across all 7 products in plain English. Quanfire answers from your live workspace — no copy-paste, no exports.
Ask in plain English — Quanfire pulls the answer from the right product.
“Show my documents and their confidence scores.”
“List this month's time entries and open matters.”
“Who are my flight-risk people right now?”
“What did my key numbers look like last quarter?”
“Show candidates for the senior associate role.”
“List my latest marketing content.”
“Give me a snapshot across my workspace.”
You stay in control — sign in with your own login, approve access, revoke anytime.
You authenticate with your own Quanfire login and approve access on a consent screen — no API keys to copy or paste around.
Out of the box your assistant reads and queries your workspace. A workspace admin can switch on write access (create, edit, delete) — and it still obeys each user's role.
Access follows your Quanfire role. Viewers stay read-only, and tenant isolation is always enforced server-side.
Your workspace data is never used to train models, and you can revoke access at any time.
The assistant only sees tools for the products your workspace actually has — nothing you're not subscribed to.
Example: Claude Desktop. Any MCP client follows the same idea — add the URL as a custom MCP server.
https://mcp.quanfire.ai/mcp.Built on the open Model Context Protocol — works with Claude and any MCP-compatible assistant.
Beyond reading: with write access enabled by your workspace admin, your assistant can also create and update records on your behalf — log time, add a candidate, update a node — always within each user's role (Viewers stay read-only).
Start a free trial, then connect Quanfire to your AI assistant in under a minute.