Figuard Mcp

io.github.figuardPublished by GitHub account figuardv1.2.0Updated Oct 24, 2026

Pre-flight spend authorization for AI agents. Set a budget, enforce limits, audit every decision.

VerifiedSTDIODesktop onlyFinanceSecurity & Monitoring

Overview

AI-generated overview

Local MCP server that gives an AI assistant FiGuard spend controls: create budgets, authorize spending before it happens, and audit every decision.

What you can do
Create and manage spending budgets, including funding, extending, resuming frozen budgets and cancelling batches of budgets, and check budget status, spent, reserved and available amounts, plus the spend ledger with filtering. Run pre-flight authorization checks that return AUTHORIZED or DENIED with a reason, then confirm the actual amount, release a failed reservation or void an unused one. Create, inspect and revoke scoped per-agent delegation tokens against a fleet budget. Run embedded against a local SQLite file, or against a self-hosted or hosted sandbox FiGuard server.
When to use it
Useful when an assistant is meant to spend money or commit resources on your behalf and you want a budget cap, a pre-flight approval step and an auditable record of each decision before anything is paid out.
Requirements
Node.js 18+. Runs as a local stdio process via npx. Embedded mode needs only a local SQLite file (default path under the home directory). Server mode needs a running FiGuard server URL in FIGUARD_BASE_URL and an API key in FIGUARD_API_KEY; a hosted sandbox is also offered. Backend mode is chosen with FIGUARD_MODE.
Before you install
The API key in FIGUARD_API_KEY is a credential and should stay in the environment, not in tool arguments. Tools create, fund, extend, cancel and void budgets and adjust limits, so they change real spend state; authorization decisions gate money. Session tokens from budget creation are short-lived but usable as tool arguments. In server or sandbox mode, budget and ledger data goes to that remote FiGuard server.

Installation

In SourceWeft

  1. Open Figuard Mcp in the dashboard and add it to a workspace.
  2. Enable the server for the chats that should use its tools.

Desktop only via STDIO. STDIO servers start a local process, so they need the SourceWeft desktop host.

Other MCP clients

Follow the launch instructions in the repository.

README

figuard-mcp

FiGuard MCP server — use FiGuard directly from Claude Code, Cursor, or Claude Desktop without writing any SDK code.

What this does

Adds 14 FiGuard tools to your AI assistant. Instead of writing SDK code, you just talk to Claude:

"Create a $500 travel budget with $300 for flights and $200 for hotels" → Claude calls figuard_create_budget

"Authorize a $149 economy flight from this budget" → Claude calls figuard_authorize → returns AUTHORIZED or DENIED

"The flight ended up costing $147 — confirm the actual amount" → Claude calls figuard_confirm

Requirements

  • Node.js 18+
  • A running FiGuard server (local or sandbox)

Setup

Claude Code

bash
claude mcp add figuard-mcp \  --env FIGUARD_API_KEY=fg_live_... \  --env FIGUARD_BASE_URL=http://localhost:8080

Or add to ~/.claude/claude_desktop_config.json manually:

json
{  "mcpServers": {    "figuard": {      "command": "npx",      "args": ["figuard-mcp"],      "env": {        "FIGUARD_API_KEY": "fg_live_...",        "FIGUARD_BASE_URL": "http://localhost:8080"      }    }  }}

Cursor

Add to .cursor/mcp.json in your project (or ~/.cursor/mcp.json globally):

json
{  "mcpServers": {    "figuard": {      "command": "npx",      "args": ["figuard-mcp"],      "env": {        "FIGUARD_API_KEY": "fg_live_...",        "FIGUARD_BASE_URL": "http://localhost:8080"      }    }  }}

Claude Desktop

Add to ~/Library/Application Support/Claude/claude_desktop_config.json (macOS):

json
{  "mcpServers": {    "figuard": {      "command": "npx",      "args": ["figuard-mcp"],      "env": {        "FIGUARD_API_KEY": "fg_live_...",        "FIGUARD_BASE_URL": "http://localhost:8080"      }    }  }}

Using the sandbox (no Docker required)

Point FIGUARD_BASE_URL at the hosted sandbox to try FiGuard without running a local server:

json
{  "FIGUARD_API_KEY": "sb_live_demo",  "FIGUARD_BASE_URL": "https://figuard-sandbox-g1ha.onrender.com"}

Available tools

ToolWhat it does
figuard_create_budgetCreate a spending envelope; returns session_token
figuard_authorizePre-flight check — AUTHORIZED or DENIED with reason
figuard_confirmClose a reservation with the actual amount after success
figuard_failRelease a reservation when the action fails
figuard_voidCancel an unused reservation
figuard_get_budgetCheck budget status, spent, reserved, available
figuard_get_ledgerList all spend events with filtering
figuard_resume_budgetUnfreeze a budget paused by anomaly detection
figuard_extend_budgetExtend a budget's expiry time
figuard_cancel_batchCancel up to 100 budgets in one call
figuard_fund_budgetAdjust a budget's total limit (CREDIT / DEBIT / RESET / RESET_SPENT)
figuard_create_delegation_tokenCreate a scoped per-agent token against a fleet budget
figuard_revoke_delegation_tokenRevoke a delegation token (idempotent)
figuard_get_delegation_tokenCheck delegation token state and cap usage

Security note

FIGUARD_API_KEY is configured as an environment variable, not a tool argument. Tool arguments appear in LLM context windows — API keys must never be passed as tool parameters.

Session tokens returned by figuard_create_budget are short-lived (they expire with the budget) and are safe to use as tool arguments within a session.

Running FiGuard locally

bash
git clone https://github.com/figuard/figuard-corecd figuard-coredocker compose up -d

FiGuard will be available at http://localhost:8080.

Source: packages/mcp/README.md at commit 9bc8430

Tools

0
SourceWeft has not started this package in its sandbox yet, so its tools are not known.

Version history

1
  1. v1.2.0LatestSep 16, 2026

Score

40% of the applicable checks met

How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.

  • Reached in the last 30 days25 pts

    Keep the endpoint answering the MCP handshake; SourceWeft probes it every 30 days.

  • Auth declared10 pts

    Declare the credential headers in the registry entry, or keep the endpoint answering the probe so its auth can be classified.

  • Tools listed15 pts

    Answer tools/list without a credential, or ship a package that starts in a sandbox without one.

  • DescriptionRequired15 pts
  • Logo10 pts
  • Documentation15 pts
  • License10 pts

    Declare a license in the repository or the registry entry.