- ✓Open-source license (MIT)
- ✓Actively maintained (<30d)
- ✓Documented (README)
- !No description
git clone https://github.com/DeepLedger/deepledger-plugin && cp deepledger-plugin/*.md ~/.claude/agents/Resumen de Subagents
# DeepLedger **AI staff accountant for QuickBooks** DeepLedger connects QuickBooks to Grok, ChatGPT, Claude and other AI agents, allowing businesses to record bills, invoices, payments, deposits, journal entries and other transactions, run financial reports, manage receivables and payables, and track month-end close using natural language. This repository packages DeepLedger as a plugin for Claude Code, Cursor, Grok Build and Grok Bot. Install it, sign in with your DeepLedger account, and ask for bookkeeping in plain language. The agent works your QuickBooks company the way a staff accountant would: it looks up the payee, checks for duplicates, picks the right transaction type, records it, and hands anything it is not sure about to a human reviewer instead of guessing. QuickBooks Online stays the ledger of record. DeepLedger holds the QuickBooks connection your company authorized through Intuit's own OAuth flow, so the plugin never sees Intuit credentials. One sign-in reaches every company you can open in the DeepLedger portal. ## Who it is for - **Small business owners and founders** who keep their own books in QuickBooks Online and want the routine entries, categorization and reports done for them. - **Bookkeepers and CPA firms** managing many client companies. One connection switches between companies, every write follows the same protocol, and uncertain items land in a review queue with the agent's reasoning attached. - **Finance teams** who want the month-end close, reconciliation prep and management reports drafted for review rather than built from scratch. ## What it does | Area | What the agent does | |------|---------------------| | Bank feed | Pulls unrecorded bank and card transactions, categorizes them from the payee's QuickBooks history, records the obvious ones and creates review tasks for the rest | | Recording | Records payments, bills, invoices, customer payments, deposits, transfers, refunds, credits and journal entries with an ID lookup and a duplicate check before every write | | Payables and receivables | Enters bills and pays them against outstanding balances, invoices customers and applies their payments, applies credits, reports aging | | Reconciliation | Matches a bank or card statement to the ledger and prepares the reconciliation workbook and open items | | Month-end close | Works the close checklist, drafts adjusting entries and the Close Sheet, and hands the package to a reviewer for sign-off | | Reports and analysis | Runs P&L, balance sheet, aging and custom reports, compares periods and explains what moved | | Audit preparation | Builds supporting schedules, an evidence index and a missing-document list | | Memory | Keeps durable per-company knowledge (policies, recurring patterns, context) and applies reviewer corrections going forward | | Human review | Anything uncertain becomes a task with the agent's reasoning; reviewer decisions are applied verbatim | ## How to use it 1. **Connect QuickBooks** at [deepledger.ai](https://deepledger.ai): Settings, QuickBooks, Connect. Repeat for each company you manage. 2. **Install the plugin** in your host (below) and sign in when the browser opens. 3. **Ask in plain language.** The `bookkeeping` skill activates on any accounting request: ```text Which QuickBooks company is active? Process the bank feed. Record: paid $500 to Office Depot for office supplies with the company credit card. Enter the Acme invoice for $2,400, net 30. Reconcile the operating account against the August statement. Generate a P&L for last month and compare it to the prior month. Close the books for June. ``` 4. **Review what it escalated.** Open Tasks in the DeepLedger portal to approve, recategorize or dismiss items the agent was not sure about. Approved tasks are recorded on the next run. The plugin itself is thin: a connector to DeepLedger's hosted MCP server and one skill. The skill does not carry procedures; it tells the agent which procedure to pull from the server's `getGuide` tool for each kind of request, so bookkeeping guidance updates on the server without a plugin release. There are no slash commands, agents, hooks or scripts. ## Prerequisites - A DeepLedger account with at least one QuickBooks Online company connected at [deepledger.ai](https://deepledger.ai) (Settings, QuickBooks, Connect) - One of the hosts below ## Install Marketplace availability depends on each marketplace's review. Until a listing appears, use the local install for that host. ### Claude Code ```bash claude plugin marketplace add DeepLedger/deepledger-plugin claude plugin install deepledger@deepledger-marketplace ``` Local development: `claude --plugin-dir ./deepledger-plugin`. ### Cursor Customize, Marketplace, search for DeepLedger, Add. Local install: clone this repository and symlink it into Cursor's local plugins folder, then reload Cursor. ```bash git clone https://github.com/DeepLedger/deepledger-plugin.git mkdir -p ~/.cursor/plugins/local ln -s "$(pwd)/deepledger-plugin" ~/.cursor/plugins/local/deepledger ``` ### Grok Build DeepLedger is listed in the [xAI plugin marketplace](https://github.com/xai-org/plugin-marketplace) as a remote source pinned to a commit of this repository. Inside Grok Build open the extensions modal with `/plugins` (or browse with `/marketplace`), select DeepLedger and install it. Grok Build reads the `.grok-plugin/plugin.json` manifest, the skill and `.mcp.json` directly. Enable and trust the plugin so Grok can connect to the bundled MCP server; the first DeepLedger tool call opens the browser sign-in, and `/mcps` shows the connection. To test before marketplace approval, install a pinned revision straight from GitHub: ```bash grok plugin install DeepLedger/deepledger-plugin@<full-commit-sha> --trust ``` Local development: `grok --plugin-dir ./deepledger-plugin`. ### Grok Bot Settings, Plugins, search for DeepLedger, Add, then complete the sign-in in your browser. Attach the plugin to a task with `@` and ask which QuickBooks company is active before requesting accounting work. ### Sign-in On first use the host discovers the server's OAuth 2.1 endpoints (authorization code with PKCE, dynamic client registration) and opens a browser sign-in. Sign in with your DeepLedger account and the connection is authorized. DeepLedger personal API keys (`dl_live_...`) are also accepted as Bearer tokens for programmatic clients. The transport is Streamable HTTP at `https://mcp.deepledger.ai/mcp`. ### Network access and credentials The plugin ships no scripts, binaries, hooks or shell commands. Everything it does goes through the hosted MCP server. The one runtime exception is the attachment upload below, where the server returns a shell-quoted `curl` command for the agent to run. The only network endpoints it reaches are: | Endpoint | Purpose | |----------|---------| | `https://mcp.deepledger.ai/mcp` | MCP server (Streamable HTTP). All QuickBooks reads and writes, guides, tasks, memory, documents and reports. | | `https://mcp.deepledger.ai/.well-known/oauth-authorization-server`, `/oauth/register`, `/oauth/authorize`, `/oauth/token`, `/oauth/revoke` | OAuth 2.1 discovery, dynamic client registration, authorization code with PKCE, token refresh and revocation. | | `https://deepledger.ai` | Browser sign-in page opened by the host during authorization. | | `https://mcp.deepledger.ai/upload-to-qb` | File attachment upload. `documents` (operation `upload`) returns a `curl` command the agent runs to send one user-chosen local file (10 MB limit) to this path with a one-time token; the server forwards it to QuickBooks. It downloads and executes nothing. | Credentials: a DeepLedger account (OAuth sign-in in the browser, scope `quickbooks`). The host stores the resulting token; the plugin never sees, stores or transmits Intuit credentials, and it reads no local files, environment variables or secrets. QuickBooks access is scoped to the companies the signed-in user can already open in the DeepLedger portal. ## How the skill works The `bookkeeping` skill activates on any accounting request. It confirms the active QuickBooks company, then calls the server's `getGuide` tool with the guide type that matches the request and follows the returned steps, safety checklist and common mistakes: | Request | `getGuide` type | |---------|-----------------| | Record a payment, bill, invoice, customer payment, refund, credit, deposit, transfer or journal entry; categorize bank feed items; AP and AR writes | `transaction_recording` | | Close a month, adjusting entries, Close Sheet | `month_end_closing` | | Reconcile a statement, explain a reconciliation difference | `reconciliation` | | Report comparisons, ratios, trends, budget variances | `financial_analysis` | | Audit or review support, evidence index, missing documents | `audit_preparation` | | A tool failure the agent cannot resolve from the response | `error_recovery` (per tool and error code) | Master data, agent memory, review tasks, documents and custom reports follow their tool descriptions; any QuickBooks write still goes through the `transaction_recording` protocol. ## What the connector provides The `.mcp.json` connector points at DeepLedger's official hosted MCP server. It installs no local binary. The server exposes 26 tools; every one operates on the active QuickBooks company and names it in its result. | Tool | Capability | |------|------------| | `qbCompanyProfile` | Show, list or switch the active QuickBooks company | | `qbMasterData` | Look up, create or update accounts, vendors, customers, items, classes and tax codes | | `qbFetchTransactions` | Fetch transactions for duplicate checks, payee history and outstanding bills or invoices | | `qbExpense` | Record or update a purchase paid now (card, ACH, check, cash) | | `qbBill` | Record or update a vendor bill to pay later | | `qbPayment` | Apply a customer payment to outstanding invoices, or pay outstandi
Lo que la gente pregunta sobre deepledger-plugin
¿Qué es DeepLedger/deepledger-plugin?
+
DeepLedger/deepledger-plugin es subagents para el ecosistema de Claude AI con 1 estrellas en GitHub.
¿Cómo se instala deepledger-plugin?
+
Puedes instalar deepledger-plugin clonando el repositorio (https://github.com/DeepLedger/deepledger-plugin) o siguiendo las instrucciones del README en GitHub. ClaudeWave también te ofrece bloques de instalación rápida en esta misma página.
¿Es seguro usar DeepLedger/deepledger-plugin?
+
Nuestro agente de seguridad ha analizado DeepLedger/deepledger-plugin y le ha asignado un Trust Score de 77/100 (tier: Trusted). Revisa el desglose completo de comprobaciones superadas y flags en esta página.
¿Quién mantiene DeepLedger/deepledger-plugin?
+
DeepLedger/deepledger-plugin es mantenido por DeepLedger. La última actividad registrada en GitHub es del 2026-10-05, con 1 issues abiertos.
¿Hay alternativas a deepledger-plugin?
+
Sí. En ClaudeWave puedes explorar subagents similares en /categories/agents, ordenados por popularidad o actividad reciente.
Despliega deepledger-plugin en tu cloud
Lleva este repo a producción en minutos. Cada plataforma genera su propio entorno con variables de entorno editables.
¿Mantienes este repo? Añade un badge a tu README
Pega el badge en tu README de GitHub para mostrar que está auditado por ClaudeWave. Cada badge enlaza de vuelta a esta página y muestra el Trust Score actual.
[](https://claudewave.com/repo/deepledger-deepledger-plugin)<a href="https://claudewave.com/repo/deepledger-deepledger-plugin"><img src="https://claudewave.com/api/badge/deepledger-deepledger-plugin" alt="Featured on ClaudeWave: DeepLedger/deepledger-plugin" width="320" height="64" /></a>Más Subagents
The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.
The agent that grows with you
Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote.
Java 面试 & 后端通用面试指南,覆盖计算机基础、数据库、分布式、高并发、系统设计与 AI 应用开发
Build Agentic workflows, RAG pipelines, with rich AI model and tool support on one collaborative workspace. Deploy on cloud, VPC, or self-hosted, so teams move from prototype to production without rebuilding the stack.
The agent engineering platform.