ContractIQ ships a first-class MCP server and a one-click Claude plugin. Every tool is permissioned, audited, and built so Claude can read, search, summarize, and propose redlines across your repository — without ever leaving your control.
ContractIQ's MCP server brokers every interaction between Claude and your repository. Tool calls inherit the user's identity and respect every existing permission boundary.
Each tool has a structured schema, scoped permissions, and a clear human-readable description that agents can reason about.
Filter by type, counterparty, status, value, dates, or a natural-language query against the full corpus.
Returns the full document tree — sections, clauses, parties, signatures — plus version history.
Cite-friendly: returns the exact text and the structural path Claude can quote back to the user.
Pull structured fields against a schema. Supports playbook-defined or ad-hoc field sets.
Agent-proposed redlines enter the same approval flow as a human's. Nothing ships without sign-off.
Returns clause-level alignment, fallback suggestions, and a Tier-1/2/3 risk classification.
Filter by owner, due date, or counterparty. Supports renewal, notice, and SLA windows.
Compose a new draft from a workspace template with structured field inputs. Routes to legal for review.
Every tool call, by every agent, with full inputs and outputs. Streamable to your SIEM.
Real conversations, real tool calls, real redlines routed back to your team.
Every guard a CLM should have, applied uniformly to humans and agents.
Annual third-party audits. Reports available under NDA.
Okta, Azure AD, Google. Group-based scope mapping for MCP tools.
Run the MCP server in your own infrastructure. No paper leaves your perimeter.
Every tool call, request, and response — signed and streamable to your SIEM.
We'll set up your workspace, walk through the tool catalog, and have your team running queries in production by end of day.
