Skip to main content
Glama
unohee

Ecount MCP Server

by unohee
OverviewSchemaRelated ServersScoreDiscussions
Python

ecount-mcp

이카운트(ECOUNT) ERP OpenAPI를 감싸는 MCP 서버. Claude 등 MCP 클라이언트에서 이카운트 ERP의 재고·판매·구매·회계 데이터를 도구로 호출한다.

구조

src/ecount_mcp/
  config.py   # 환경변수 → EcountConfig (인증 정보, 테스트/운영 도메인)
  client.py   # Zone → Login(SESSION_ID) → Data API 인증 흐름 + REST 호출
  server.py   # FastMCP 서버. @mcp.tool 로 엔드포인트 노출
docs/
  ecount-openapi.md   # 이카운트 OpenAPI 인증/엔드포인트 구조 정리

Related MCP server: ezPay E-Invoice MCP Server

설치

python -m venv .venv && source .venv/bin/activate
pip install -e ".[dev]"

설정

.env.example → .env 로 복사 후 채운다 (또는 MCP 클라이언트 env로 주입):

ECOUNT_COM_CODE=...
ECOUNT_USER_ID=...
ECOUNT_API_CERT_KEY=...
ECOUNT_DOMAIN=sboapi   # 테스트=sboapi, 운영=oapi

API 인증키는 이카운트 ERP 관리자 화면에서 발급한다(테스트키/운영키 구분).

실행

ecount-mcp            # stdio 트랜스포트 (Claude Desktop 등)

또는 모듈 직접 실행:

python -m ecount_mcp.server

Claude Desktop 등록 예

{
  "mcpServers": {
    "ecount": {
      "command": "ecount-mcp",
      "env": {
        "ECOUNT_COM_CODE": "...",
        "ECOUNT_USER_ID": "...",
        "ECOUNT_API_CERT_KEY": "...",
        "ECOUNT_DOMAIN": "sboapi"
      }
    }
  }
}

도구

도구

설명

ecount_call(endpoint, payload)

임의의 OAPI/V2/... 엔드포인트 호출

ecount_save_customer(customers)

거래처등록 (매뉴얼 §4.1)

ecount_save_invoice(invoices)

매출·매입전표 II 자동분개 (매뉴얼 §9) — 세금계산서 발행 직전 분개

ecount_web_export_tax_invoices()

전자(세금)계산서 리스트(매입) 웹 세션 추출 → 행(JSON). OAPI 미지원, 헤드리스 무인

저장 도구는 부분실패를 처리한다: 반환값 {success_cnt, fail_cnt, slip_nos, failures[], all_ok}. ecount_save_invoice는 호출 전 입력검증(매출=CR_CODE, 매입=DR_CODE 필수)으로 rate limit을 보호한다.

웹 세션 추출 (ecount_web_export_tax_invoices) — OAPI 미지원 보완

OAPI에 없는 화면(전자세금계산서 리스트 등)은 인증된 웹 세션 + Playwright로 추출한다([web] extra).

pip install -e ".[web]" && playwright install chromium

  • 무인 헤드리스 로그인: 신뢰기기 프로필(ECOUNT_WEB_PROFILE_DIR) + ECOUNT_COM_CODE/ECOUNT_USER_ID/ECOUNT_USER_PW 환경변수로 자동 로그인(2FA 생략).

  • zone 지정 (필수): 웹 호스트는 login{zone}.ecount.com. 자기 회사 zone을 ECOUNT_WEB_ZONE에 넣는다 — 기본값은 없다. zone 값은 OAPI Zone API 반환값과 같다.

  • 메뉴 해시: 화면 진입 경로는 회사별 메뉴 구성에 따라 다를 수 있다. 기본값이 안 맞으면 ECOUNT_WEB_TAX_MENU_HASH로 덮어쓴다(브라우저 주소창의 #... 부분).

  • 선행 1회(부트스트랩): 사람이 신뢰기기를 등록해야 한다 — python testing/hometax_invoice_export_260612_v3.py --run --headed 로 1회 GUI 로그인(2FA 통과). 이후 그 프로필로 무인 추출이 된다. 신뢰기기 만료 시 재부트스트랩.

  • 반환: {columns, rows(dict 리스트), count}. 인증 산출물(.ecount_profile/·*.xlsx)은 .gitignore.

세금계산서 워크플로우 (이 프로젝트의 핵심 목적)

[자동: MCP]                                    [수동: 사람]
거래처 확인/등록 → ecount_save_invoice(분개) → ERP에서 전표 검수 → 발행 버튼
(ecount_save_customer)   ↑ 전표번호 발급         ↑ 전자세금계산서 발행은 OAPI 범위 밖

  • 분개(ecount_save_invoice)까지가 MCP 범위. 매출: TAX_GUBUN="11" + CR_CODE(예 "4019" 상품매출), 매입: TAX_GUBUN="21" + DR_CODE(예 "1469" 상품). + CUST, SUPPLY_AMT, VAT_AMT.

  • 전자세금계산서 발행은 사람이 ERP에서 검수 후 직접 (의도된 검수 게이트 — 발행=법적효력).

  • CUST/CR_CODE/DR_CODE/TAX_GUBUN 코드값은 OAPI로 조회 불가(매뉴얼에 코드 조회 API 없음) → ERP에 등록된 값을 사용. 상세: docs/ecount-openapi.md ★섹션.

새 엔드포인트는 src/ecount_mcp/server.py 에 @mcp.tool 로 얇게 추가한다.

  • 전체 API 명세: 이카운트 공식 OpenAPI 매뉴얼 (벤더 저작물이라 이 저장소에 포함하지 않는다. ERP 로그인 후 자기설정 > OpenAPI 에서 받아 docs/ecount-api-manual.md 로 두면 아래 문서들의 참조가 맞는다)

  • 코드 매핑·함정: docs/ecount-openapi.md

상태

핵심 목표(세금계산서 분개 자동화) 실서버 검증 완료 (2026-06-02, 테스트존 sboapi). end-to-end 실호출로 확인:

  • Zone → Login(SESSION_ID) → 거래처등록 → 매출분개 성공(전표번호 발급)

  • stdio MCP 핸드셰이크(initialize/tools/list) 정상, 입력검증·부분실패 파싱 동작

  • 단위 테스트 16 passed

나머지 엔드포인트(품목/영업/구매/생산/재고/쇼핑몰/근태/게시판)는 매뉴얼 편입 완료, 도구화는 필요 시 추가 (docs/ecount-openapi.md TODO). 일부 매뉴얼 Example은 본문 잘림으로 보강 대기.

구현 스펙·자동화 가능/불가 범위 전체 정리: docs/IMPLEMENTATION-STATUS.md (구현된 도구, OAPI로 가능한 미도구화 엔드포인트, 구조적으로 불가능한 부분, 운영 제약, 다음 작업 후보)

Available Tools

4 tools
ecount_callC

이카운트 OAPI/V2 엔드포인트를 호출한다.

ParametersJSON Schema
NameRequiredDescriptionDefault
payloadNo요청 본문(JSON). 엔드포인트별 파라미터.
endpointYes"/OAPI/V2/" 뒤의 경로. 예: "InventoryBalance/GetListInventoryBalanceStatusByLocation"

Output Schema

ParametersJSON Schema
NameRequiredDescription

No output parameters

TDQS

C2.2/5.0
Behavior1/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

No annotations are provided, so the description carries the full burden of behavioral disclosure. It fails to indicate whether the call is read-only or mutating, authentication requirements, rate limits, or side effects. The description only says 'calls the endpoint,' giving zero behavioral context.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness2/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is a single short sentence, but it is under-specified rather than concise. It fails to convey essential context, so the sentence does not 'earn its place.' Similar to the 'Process' example, this is a lack of content, not effective conciseness.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness2/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a generic OAPI/V2 caller with no annotations, the description is insufficient. It does not explain that behavior depends on the endpoint, potential side effects, or error handling. Although the schema covers parameters and an output schema exists, the description does not provide enough context for safe agent invocation.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

The input schema has 100% description coverage for both parameters (endpoint and payload). The tool description itself adds no parameter information beyond what the schema provides. Since schema coverage is high, the baseline is 3, and there is no additional value from the description.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose3/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description identifies the action ('calls') and resource ('ecount OAPI/V2 endpoint') but is vague about what the endpoint actually does. It does not differentiate this generic call tool from sibling tools like ecount_save_customer or ecount_save_invoice, which are specific operations. It is not a tautology, but it lacks specificity.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines2/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description provides no information about when to use this tool versus the sibling tools. It does not state that this is a low-level generic caller for arbitrary endpoints, nor does it mention any exclusions or alternatives. An agent has no guidance on tool selection.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

ecount_save_customerB

거래처를 등록한다 (매뉴얼 §4.1 AccountBasic/SaveBasicCust).

ParametersJSON Schema
NameRequiredDescriptionDefault
customersYes거래처 dict 리스트. 각 항목 필수 키: - BUSINESS_NO: 사업자등록번호(거래처코드) - CUST_NAME: 회사명 그 외 BOSS_NAME, TEL, EMAIL 등 선택 항목은 매뉴얼 §4.1 참조. (입력 가능 항목은 ERP 회사코드별 기본탭 설정에 따라 다름)

Output Schema

ParametersJSON Schema
NameRequiredDescription

No output parameters

TDQS

B3.2/5.0
Behavior2/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations provided, the description carries the full burden of behavioral disclosure. '거래처를 등록한다' implies a write operation (creating a customer), but there's no information about duplicate handling, required permissions, side effects, or response format. The manual reference is a pointer, not an actual disclosure.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is a single, concise sentence that conveys the core purpose and includes a manual reference for details. Every word earns its place; no fluff or redundancy.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness2/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

The tool is a mutation (write) operation with no annotations and a minimal description. It doesn't address potential complexities like duplicate customers or ERP-specific settings, leaving the AI agent under-informed. The output schema exists but is not visible, and the description does not compensate for the lack of behavioral context.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

The schema description covers 100% of the only parameter (customers), detailing required and optional keys. The tool description itself adds no parameter-level information, but since schema coverage is high, a baseline of 3 is appropriate.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the tool's function: '거래처를 등록한다' (register a customer), which is a specific verb and resource, and is distinct from sibling tools like ecount_save_invoice or ecount_web_export_tax_invoices. The manual reference adds credibility and precision.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines2/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

No explicit guidance on when to use this tool versus alternatives. While it's obvious that it's for customer registration, the description doesn't mention scenarios, prerequisites, or exclusions. The manual reference could imply more details, but it's not stated in the description itself.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

ecount_save_invoiceA

매출·매입전표 II 자동분개를 입력한다 (매뉴얼 §9 InvoiceAuto/SaveInvoiceAuto).

세금계산서 '발행 직전' 회계 분개까지를 자동화한다. 실제 전자세금계산서 발행은 이카운트 ERP에서 사람이 검수 후 직접 처리한다(OAPI 범위 밖).

ParametersJSON Schema
NameRequiredDescriptionDefault
invoicesYes전표 dict 리스트. 각 항목 주요 키: - TAX_GUBUN (필수): 매출/매입 부가세유형 코드. 예) "11"=매출, "21"=매입. - CR_CODE: 매출 시 매출계정코드 (예 "4019" 상품매출). 매출이면 필수. - DR_CODE: 매입 시 매입계정코드 (예 "1469" 상품). 매입이면 필수. - CUST: 거래처코드, SUPPLY_AMT: 공급가액, VAT_AMT: 부가세. - TRX_DATE: 전표일자(YYYYMMDD, 미입력 시 현재일), REMARKS: 적요 등. 전체 항목은 매뉴얼 §9 참조.

Output Schema

ParametersJSON Schema
NameRequiredDescription

No output parameters

TDQS

A4/5.0
Behavior3/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations, the description carries full responsibility for behavioral disclosure. It reveals the most critical trait: this tool does NOT issue electronic tax invoices, only prepares the journal entry. This is valuable context. However, it does not mention side effects, idempotency, validation behavior, permissions, or error handling, leaving gaps in transparency beyond the single boundary condition.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is exceptionally concise: two sentences that immediately state the action, scope, and critical limitation. It is front-loaded with the primary purpose, includes a useful manual reference, and contains no redundant or filler content.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given the tool's moderate complexity (one nested array parameter) and the presence of an output schema, the description adequately covers the most crucial contextual boundary: the distinction between journal entry preparation and actual tax invoice issuance. It also points to the manual for full parameter details. However, it omits operational prerequisites or failure behavior, which would push it to a 5.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

The input schema already provides 100% coverage with detailed descriptions of the 'invoices' array items, including required keys like TAX_GUBUN, CR_CODE/DR_CODE, and optional fields like TRX_DATE and REMARKS. The tool description adds only a manual reference ($9) and does not supplement parameter meaning beyond the schema. Hence, baseline 3 applies.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the tool's function: 'Enters sales/purchase slip II automatic journal entries' (매출·매입전표 II 자동분개를 입력한다). It also distinguishes itself from siblings by clarifying that actual e-tax invoice issuance is outside OAPI scope, implying ecount_web_export_tax_invoices or human action covers that, while ecount_save_customer is for customers. This is a specific verb+resource with clear differentiation.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines4/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description provides clear context on when to use: for automating accounting journal entries up to but not including the actual tax invoice issuance. It explicitly states the tool's limitation ('actual e-tax invoice issuance is handled by a person after review in ecount ERP (outside OAPI scope)'), which helps avoid misuse. However, it does not explicitly name alternative sibling tools or provide step-by-step usage conditions, so it falls slightly short of a 5.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

ecount_web_export_tax_invoicesA

전자(세금)계산서 리스트(매입)를 웹 세션으로 추출한다 (OAPI 미지원, 화면 ECTAX102M).

OAPI에는 이 홈택스 수집자료 조회 기능이 없어, 인증된 웹 세션으로 화면을 열고 엑셀을 받아 행(JSON)으로 돌려준다. 헤드리스 무인 로그인(신뢰기기 프로필 + ECOUNT_COM_CODE/USER_ID/USER_PW 환경변수). 상세: docs/IMPLEMENTATION-STATUS.md, KT-274.

Returns: {columns: 컬럼명 리스트, rows: 행 dict 리스트(컬럼명→값), count: 건수}. 신뢰기기 프로필이 없거나 2FA가 요구되면 RuntimeError(부트스트랩 안내).

Note: 선행 1회: 사람이 python testing/hometax_invoice_export_260612_v3.py --run --headed로 로그인해 기기를 등록해야 이후 무인 추출이 된다(신뢰기기 등록).

ParametersJSON Schema
NameRequiredDescriptionDefault

No parameters

Output Schema

ParametersJSON Schema
NameRequiredDescription

No output parameters

TDQS

A4.5/5.0
Behavior5/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations, the description carries full burden and does an excellent job. It discloses the OAPI limitation, web-session-based behavior, headless login via env vars, return format (columns/rows/count), failure conditions (RuntimeError on missing trusted device or 2FA), and the required first-time manual setup. This is comprehensive.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is moderately detailed but well-structured with clear sections (main description, returns, note). Each sentence provides value, including the reference to implementation docs and the bootstrap step. Slightly verbose but appropriate for the complexity.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a no-parameter tool, the description covers the purpose, return structure, prerequisites, and failure modes. It also references additional documentation. It could be slightly more explicit about what the column/row data contains, but given an output schema exists, this is sufficient.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

The input schema has zero parameters, so the baseline is 4. The description mentions environment variables used for login, but these are not tool parameters and are contextually useful. No parameter confusion exists.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states that it extracts a purchase electronic tax invoice list via a web session, with a specific screen (ECTAX102M). It distinguishes itself from siblings (save/call tools) by focusing on export/read, and mentions an OAPI limitation that makes this tool necessary.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines4/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

It explains when to use the tool (because OAPI does not support this feature) and provides a prerequisite (one-time manual login for trusted device registration). However, it does not explicitly name alternative tools or give exclusions, but the context is clear enough.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections. Dates show when Glama detected each change.

  1. 4 tool updatesv0.1.0
    • First observedecount_call
    • First observedecount_save_customer
    • First observedecount_save_invoice
    • First observedecount_web_export_tax_invoices
Install Server
RayrunBuild, deploy, and operate MCP servers without becoming an infrastructure team.ray.runAd

TDQS

B3.4/5.0
Disambiguation5/5

Each tool has a clearly distinct purpose: ecount_call is a generic API caller, while ecount_save_customer, ecount_save_invoice, and ecount_web_export_tax_invoices handle specific operations. There is no overlap between saving customers, saving invoices, and exporting tax invoices.

Naming Consistency3/5

Naming is mixed: 'ecount_save_customer' and 'ecount_save_invoice' follow a verb_noun pattern, but 'ecount_call' is just a verb and 'ecount_web_export_tax_invoices' includes a modifier and longer verb phrase. The inconsistency reduces predictability.

Tool Count5/5

With only 4 tools, the server is well-scoped for its purpose of handling customers, invoices, and tax invoice export. Each tool earns its place without redundancy or overload.

Completeness4/5

The server covers the core workflow of saving customer and invoice data and extracting tax invoices. The generic ecount_call can access other endpoints, mitigating gaps, but dedicated retrieval or update tools for customers/invoices are absent.

Maintenance

ActivityMaintained
ResponsivenessSyncing

Resources

Unclaimed servers have limited discoverability.

Looking for Admin?

If you are the server author, to access and configure the admin panel.

Related MCP Connectors

Related MCP Servers

  • MCP4Acumatica

    hallboys
    A
    license
    Not graded
    quality
    A
    maintenance
    Enables Claude to interact with Acumatica ERP through a remote MCP server with per-user OAuth, role-based access, and 44 tools for querying and managing ERP data.
    17
    Apache 2.0

  • ECOUNT MCP Server

    storehausai
    A
    license
    A
    quality
    D
    maintenance
    Enables AI assistants to interact with ECOUNT ERP through natural language, providing tools for products, inventory, sales, purchases, and more.
    23
    28
    3
    MIT

  • shopline-mcp

    tzangms
    A
    license
    A
    quality
    B
    maintenance
    MCP server for the Shopline Open API. Exposes 140+ tools for querying and managing orders, products, customers, promotions, analytics, and store settings from your Shopline store via Claude.
    100
    MIT

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/unohee/ecount-mcp'

If you have feedback or need assistance with the MCP directory API, please join our Discord server