Tmp.BZxIpFB2PM

creditkarma-mcp

MCP server for Credit Karma — 同步s transactions into a local SQLite database and provides natural-language 查询ing 工具s.

npm: npmjs.com/package/creditkarma-mcp Source: github.com/chrischall/creditkarma-mcp 设置up Option A — npx (recommended)

添加 to .mcp.json in your project or ~/.claude/mcp.json:

{ "mcpServers": { "creditkarma": { "command": "npx", "args": ["-y", "creditkarma-mcp"], "env": { "CK_COOKIES": "your-ckat-value-here" } } } }

Option B — from source git clone https://github.com/chrischall/creditkarma-mcp cd creditkarma-mcp npm 安装 && npm 运行 build

Then 添加 to .mcp.json:

{ "mcpServers": { "creditkarma": { "command": "node", "args": ["/path/to/creditkarma-mcp/dist/索引.js"], "env": { "CK_COOKIES": "your-ckat-value-here" } } } }

Or use a .env file in the project directory with CK_COOKIES=.

获取ting CK_COOKIES

Scripted (recommended — source 安装):

npm 运行 auth # prints the CKAT value to the console npm 运行 auth -- .env # writes CK_COOKIES= to .env

Launches Chrome with a dedicated 性能分析, wAIts for 签名-in at creditkarma.com, then captures the CKAT cookie (the URL-encoded bundle of 访问 + refresh JWTs). Use the printed value with Claude 桌面 / MCPB, or the .env form when 运行ning from source.

Manual (Dev工具s):

记录 in to creditkarma.com in Chrome Dev工具s → 应用 → Cookies → creditkarma.com Copy the CKAT cookie value

Accepts: raw CKAT value, CKAT=, or the full Cookie header string from any CK network 请求.

Authentication

Call ck_设置_会话 with your cookie value to store 凭证s and enable auto-refresh.

访问 令牌: ~15 min TTL, auto-refreshed transparently Refresh 令牌: ~8 hours TTL When expired: re-运行 npm 运行 auth (or grab a new CKAT cookie) and call ck_设置_会话 工具s Auth 工具 Description ck_设置_会话(cookies) Store 凭证s — accepts CKAT value, CKAT=, or full Cookie header 同步 工具 Description ck_同步_transactions(force_full?) 同步 transactions to local SQLite. Incremental by default (since last 同步 − 30 days). force_full=true re-fetches everything. 查询 工具 Description ck_列出_transactions(启动_date?, end_date?, account?, category?, merchant?, 状态?, min_amount?, max_amount?, limit?, off设置?) 过滤器ed, paginated transaction 列出 ck_获取_recent_transactions(limit?) N most recent transactions (default 20) ck_获取_spending_by_category(启动_date?, end_date?) Spending totals grouped by category ck_获取_spending_by_merchant(启动_date?, end_date?, limit?) Spending totals grouped by merchant ck_获取_account_summary Transaction counts and totals per account ck_查询_sql(sql) Read-only SQL 查询 agAInst the local database (SELECT only) 工作流s

First-time 设置up:

运行 npm 运行 auth (or grab the CKAT cookie manually from creditkarma.com Dev工具s) Paste into CK_COOKIES env var, or call ck_设置_会话(cookies) from within Claude ck_同步_transactions → initial full 同步

Regular use:

ck_同步_transactions → pull latest transactions Then 查询 with any of the 查询 工具s

Spending analysis:

ck_同步_transactions ck_获取_spending_by_category(启动_date: "2026-01-01", end_date: "2026-03-31") ck_获取_spending_by_merchant(启动_date: "2026-01-01", limit: 10)

Custom analysis with SQL:

-- Monthly spending totals SELECT strftime('%Y-%m', date) AS month, SUM(ABS(amount)) AS total FROM transactions WHERE amount < 0 GROUP BY month ORDER BY month DESC

-- Spending by category this year SELECT c.name, SUM(ABS(t.amount)) AS total FROM transactions t JOIN categories c ON t.category_id = c.id WHERE t.date >= '2026-01-01' AND t.amount < 0 GROUP BY c.name ORDER BY total DESC

Database 模式 transactions (id, date, description, 状态, amount, account_id, category_id, merchant_id, raw_json) accounts (id, name, type, 提供者_name, display) categories (id, name, type) merchants (id, name) 同步_状态 (key, value)

Notes All 查询 工具s 运行 agAInst the local SQLite database — 同步 first Amounts: negative = expense/debit, positive = credit/income ck_查询_sql only allows SELECT — no writes to Credit Karma data 同步 saves a 恢复 cursor — interrupted 同步s can be 恢复d automatically