Candor Finance

The package is source-available, so an agent platform or reviewer can inspect the declared server, repository, finance skill, and connection metadata before installation.

An authorized agent gets a complete financial workspace: accounts, balances, transactions, recurring items, budgets, goals, holdings, debts, notes, and factual changes, each carrying freshness, coverage, and evidence handles.

An authorized agent can work with normalized accounts, balances, transactions, recurring items, budgets, goals, holdings, debts, notes, factual changes, and prior financial-impact records. Responses preserve freshness, coverage, caveats, and evidence handles so the agent can distinguish what the records establish from what still needs judgment.

The tool surface is designed for progressive disclosure. The agent opens the workspace, asks for the schema or bounded records a task needs, and keeps large or unrelated financial data out of the conversation.

Why the finance skill accompanies the connection

The MCP server transports tools and data. The candor-finance skill teaches the agent the working methods, and supported installs pair both at matching versions.

MCP or the CLI transports tools and data. The matching candor-finance skill teaches the agent how to do the work well: checking coverage, investigating a possible duplicate charge, maintaining an approved budget, recording evidence, or following up on a missing refund. The supported package keeps the method and tool contract on the same version.

Start Candor gives the user's agent verified skill, MCP, CLI, package, and harness-specific materials. The agent recognizes its own environment and assembles a complete supported route without treating a bare server connection as a complete installation.

Access uses OAuth 2.1 with dynamic client registration. You approve each client on Candor's secure page, and bank credentials never pass through the agent or the chat.

You can revoke an OAuth client, disconnect a financial account, export your organized data, or delete what Candor holds from the account controls.

The server has no tools that move money or change anything at a bank or broker. Writes are limited to workspace state you approve, such as budgets, goals, and notes.

Candor cannot transfer money, pay a bill, execute a trade, open or close an account, cancel a subscription, file a document, or contact another party. Those tools do not exist in the MCP server.

Inside the workspace, the agent can make scoped, auditable updates such as an approved budget, goal, correction, or its own follow-through note. Choices that reflect your values require your approval. Any action outside Candor belongs to your agent under the authority you grant it.

A Model Context Protocol server that gives an AI agent structured tools for financial data. Candor Finance is a hosted one: the agent connects over remote Streamable HTTP with OAuth and works with your connected accounts, spending, budgets, goals, and records.

Any MCP client that supports remote Streamable HTTP servers with OAuth, including Claude, OpenClaw, Hermes, and ChatGPT. Agents without MCP support can use the Candor CLI instead.

Is it safe to connect an AI agent to my bank accounts?

The connection is read-only toward your bank. Accounts link through a bank-grade provider flow, your bank login stays with the provider, and the server has no tools that move money. Every read is logged, and you can revoke access anytime.

How much does the Candor Finance MCP server cost?

The server is included in a Candor subscription. Current pricing and trial terms are shown before checkout on candor.money.

The canonical record is money.candor/candor-finance on the official MCP Registry. It points at the production endpoint api.candor.money/mcp and the public source package on GitHub.

One message to the agent you already use. You see the price before anything private, approve the access and accounts yourself, and your agent is working within minutes.

Bridge Town is an MCP-native, git-versioned financial modeling platform for FP&A teams and finance leaders. AI agents use Bridge Town tools to create projects, write Python model files, run models in isolated cloud sandboxes, query data, write outputs to Google Sheets, create dashboards, branch scenarios, and collaborate with teammates.

The Capital.com MCP Server lets your AI assistant talk to your trading account directly. Market data, position checks, trade previews – all in plain language, without leaving your AI tool.

Coinrule Agentic Trading MCP enables investors to create, backtest, execute, and manage trading agents through natural language across stocks, crypto and ETFs

Invest with Claude and other AI assistants

Australian Consumer Data Right Product Data

Remote MCP server for historical crypto & prediction-market data: search ~500K instruments, live market stats (OHLC, turnover, spreads, depth, slippage) and tick-data purchase. Keyless for catalog & stats; optional OAuth for account tools. Endpoint: https://cryptostruct.com/mcp

Cross-border debt collection from your AI assistant: check cases, get pricing, submit new cases.

Read-only MCP server for your Evibe investment portfolio + live market data (holdings, performance, dividends, benchmarks, screeners). Works with Claude & ChatGPT.