avie-moneylover-mcp
Provides tools to read Money Lover account data and add transactions, with features for previewing, duplicate detection, reconciliation, and safe batch import of bank/card statement JSON via an MCP client.
README
Avie Money Lover MCP / Local Bridge
Unofficial Money Lover integration for two workflows:
- MCP server — expose Money Lover read/write tools to an MCP-capable client.
- GPT + local bridge — let ChatGPT prepare/review statement JSON while your own machine keeps the Money Lover token/cookie and performs the final import.
Unofficial project. This repository is not affiliated with, endorsed by, or sponsored by Money Lover. It uses web endpoints observed from the Money Lover web app. Those endpoints, authentication flows, category behavior, and anti-bot requirements can change without notice.
What it can do
- Read authenticated user info
- List wallets and wallet categories
- Read transactions by date range
- Resolve Money Lover
user_category_v2source IDs to runtime IDs using transaction history - Preview a single transaction without writing
- Add a transaction with duplicate protection
- Preview a bank/card statement batch
- Reconcile expense − income against a statement total
- Flag exact duplicates, probable duplicates, unresolved categories, and review rows
- Import sequentially only after explicit confirmation
- Verify created transactions by reading them back
- Safely re-run after a partial import: exact duplicates are skipped
Requirements
- Windows 10/11, macOS, or Linux
- Node.js 22+
- A Money Lover account you control
- For reliable writes: a fresh wallet-owner session, current browser cookie/
cf_clearanceif required, and the matching browser User-Agent
Quick start
git clone https://github.com/haonguyen97ftu/avie-moneylover-mcp.git
cd avie-moneylover-mcp
npm install
npm test
Windows helper setup
scripts\setup-windows.cmd
Set credentials in the current shell only:
set "MONEYLOVER_ACCESS_TOKEN=YOUR_FRESH_OWNER_TOKEN"
set "MONEYLOVER_CF_CLEARANCE=YOUR_CURRENT_CF_CLEARANCE_VALUE"
set "MONEYLOVER_USER_AGENT=YOUR_EXACT_BROWSER_USER_AGENT"
Then run:
scripts\doctor.cmd
Recommended GPT + local workflow
Bank/card statement
↓
ChatGPT extracts + categorizes
↓
statement JSON
↓
local preview
↓
review / duplicate / reconciliation checks
↓
explicit IMPORT confirmation
↓
Money Lover
↓
local verification report
Do not paste Money Lover tokens, refresh tokens, passwords, browser cookies, or cf_clearance into ChatGPT, GitHub issues, logs, or screenshots. ChatGPT only needs the statement plus the generated preview/result JSON.
Preview a statement
scripts\preview.cmd data\your-statement.json
Import after review
scripts\import.cmd data\your-statement.json
The importer requires you to type exactly:
IMPORT
before writes begin.
See docs/HUONG_DAN_SU_DUNG.md for the complete Vietnamese guide.
API behavior observed
Authenticated requests currently use:
Authorization: AuthJWT <access_token>
Observed endpoints:
| Client method | HTTP | Endpoint |
|---|---|---|
getUserInfo() |
POST | /api/user/info |
getWallets() |
POST | /api/wallet/list |
getCategories(walletId) |
POST | /api/category/list |
getTransactions(...) |
POST | /api/transaction/list |
addTransaction(...) |
POST | /api/transaction/add |
Writes may require browser-like headers and a current Cloudflare session. Supported environment variables include:
MONEYLOVER_ACCESS_TOKENMONEYLOVER_CF_CLEARANCEMONEYLOVER_COOKIEMONEYLOVER_USER_AGENTMONEYLOVER_TIMEOUT_MSMONEYLOVER_CATEGORY_LOOKBACK_DAYSMONEYLOVER_WRITE_DELAY_MS
Batch statement format
Start from data/example-statement.json. Full field documentation is in docs/STATEMENT_FORMAT.md.
Preview:
npm run preview -- data\your-statement.json
Import:
npm run import -- data\your-statement.json
Optional flags:
--allow-review— include rows intentionally left inreviewafter you inspect them--allow-mismatch— bypass statement reconciliation protection; use only when you understand the discrepancy--allow-shared-wallet— bypass the owner safety check; shared-wallet writes may still fail--confirm IMPORT— non-interactive confirmation; not recommended for ordinary manual use
Safety defaults
The importer refuses to write when:
- any row is
blocked - a
reviewrow exists unless--allow-reviewis supplied - the statement reconciliation difference is non-zero unless
--allow-mismatchis supplied - the authenticated user is not the wallet owner unless
--allow-shared-walletis supplied - confirmation is not exactly
IMPORT
Amounts must be positive numbers.
Category v2
Some accounts expose a source/template category ID from /category/list, while /transaction/add expects a different runtime category ID. The bridge learns this mapping from existing transactions.
If a category has never appeared in history, copy:
config/runtime-category-overrides.example.json
to:
config/runtime-category-overrides.json
and add the runtime ID observed from a real browser transaction request. The local override file is ignored by Git.
Public-repo safety
The repository is configured to avoid committing local financial data by default:
data/*.jsonis ignored except the sanitized exampleout/is ignored.env*is ignored- local runtime category overrides are ignored
- local session helper files are ignored
Before pushing, run:
npm run repo:check
npm test
Then inspect git status manually.
Security
If a token, password, refresh token, browser cookie, or cf_clearance value is ever exposed, treat it as compromised and rotate/revoke the session. See SECURITY.md.
Contributing
Bug reports and pull requests are welcome. Please read CONTRIBUTING.md before submitting changes. Never include real financial statements, account identifiers, tokens, or cookies in issues or test fixtures.
Documentation
- Vietnamese usage guide
- Statement JSON format
- Architecture
- GitHub publishing guide
- Security policy
- Changelog
- GPT statement normalization prompt
- GPT preview review prompt
License
MIT — see LICENSE.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.