barcoding-mcp
Enables encoding 100+ barcode symbologies and decoding common 1D/2D formats, plus terminal QR codes.
README
barcoding-mcp
An MCP server that encodes 100+ barcode symbologies (via bwip-js) and
decodes all the common 1D/2D formats (via zxing-wasm), with zero native
dependencies. Node runtime, served over both stdio and streamable HTTP.
Plus an ASCII/Unicode terminal QR for output that needs no image channel.
The funnel is asymmetric on purpose: encode 100+, decode the common set, round-trip-verified where both overlap. We do not claim "decode anything."
| Direction | Library | Coverage |
|---|---|---|
| Encode | bwip-js |
100+ symbologies (full BWIPP) |
| Decode | zxing-wasm |
~20 common 1D/2D (QR, Micro QR, DataMatrix, Aztec, PDF417, MaxiCode, Code128/39/93, EAN/UPC, ITF, Codabar, DataBar) |
| Terminal QR | qrcode |
block-character QR, pure text |
Tools
| Tool | Purpose |
|---|---|
encode_barcode |
Render a barcode (bcid + text + typed common options + freeform options bag). Returns a PNG. |
encode_qr_terminal |
Encode data to a block-character QR. style: unicode (default, safe in codeblocks/logs) or ansi (live TTY). |
decode_barcode |
Detect + decode 1D/2D barcodes from an image (base64 / path / url; PNG or JPEG). |
decode_batch |
Decode many images in one call; per-item results, isolated errors. |
decode_pdf |
Rasterize a PDF (mupdf WASM) and pull every barcode, tagged with page number. |
gs1_parse |
Parse a decoded GS1 string into structured Application Identifiers (GTIN, dates, batch, serial, weights…). |
verify_barcode |
Encode → decode the render → assert the payload round-trips. Self-verifying label QA. |
list_symbologies |
Supported symbologies flagged encode / decode / both. |
list_symbology_options |
Valid encode options for a bcid, segmented specific / common / sizing. Discover the options bag's keys before encoding. |
Install & run
npm (stdio — for Claude Code, Cursor, local agents)
npx @cordfuse/barcoding-mcp # stdio transport (default)
npx @cordfuse/barcoding-mcp --http # streamable HTTP on :3900
MCP client config (stdio):
{
"mcpServers": {
"barcoding": { "command": "npx", "args": ["-y", "@cordfuse/barcoding-mcp"] }
}
}
Docker / GHCR (streamable HTTP — for remote / metamcp wiring)
docker run -p 3900:3900 ghcr.io/cordfuse/barcoding-mcp:latest
# or, from a checkout:
docker compose -f docker/compose.yaml up
Serves streamable HTTP at http://<host>:3900/mcp, with a liveness probe at
/health. Port is settable via the PORT env. stdio is per-invocation and not
containerised — the image is the HTTP endpoint.
Repository layout (monorepo)
barcoding-mcp/
packages/
mcp/ @cordfuse/barcoding-mcp (the server) → packages/mcp/README.md
docker/ Dockerfile + compose.yaml (the --http server)
.github/ CI + release workflows
npm workspace; node_modules hoists to the root.
Development
npm ci # install (root)
npm run build -w @cordfuse/barcoding-mcp # tsc + copy option catalog into dist
npm test -w @cordfuse/barcoding-mcp # node:test smoke suite
npm run gen:options -w @cordfuse/barcoding-mcp # regenerate the option catalog from bwip-js
The list_symbology_options catalog is generated from bwip-js's own sources
(dist/bwip-js-gen.d.ts for common options, barcode.ps for the per-symbology
option preambles) into packages/mcp/src/data/symbology-options.json.
CI / Release
- CI (
.github/workflows/ci.yml) — typecheck + build + test on Node 20 & 22 for every push and PR tomain. - Release (
.github/workflows/release.yml) — pushing av*tag publishes:- the npm package
@cordfuse/barcoding-mcp(guards that the tag matches the package version), and - the Docker image
ghcr.io/cordfuse/barcoding-mcp:<version>+:latest.
- the npm package
npm version patch -w @cordfuse/barcoding-mcp
git commit -am "release vX.Y.Z" && git tag vX.Y.Z && git push --follow-tags
Secrets: npm publish needs the org NPM_TOKEN (this repo must be on the
org token's selected-repo allow-list). GHCR uses the built-in GITHUB_TOKEN
(packages: write) — no secret required.
License
MIT
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。