Unofficial LLM-ready knowledge base for Grow / Meshulam (גרואו · משולם) payments — the Light API as llms.txt + MCP server + merged OpenAPI & Postman. Not affiliated with Grow/Meshulam.
Grow Payments — LLM-ready API knowledge base (llms.txt + MCP)
Grow (formerly & still widely known as Meshulam) · גרואו · משולם
🌐 Bilingual content — English & Hebrew (עברית) · 🇮🇱 Israeli payment gateway
⚠️ UNOFFICIAL — WE ARE NOT GROW / Meshulam. Not affiliated. Not endorsed.
בסיס ידע לא רשמי לתיעוד המפתחים של Grow / משולם — מותאם ל‑LLM: קורפוס llms.txt ושרת MCP.
התוכן דו‑לשוני (אנגלית + עברית). זהו פרויקט עצמאי — איננו Grow / משולם ואיננו מסונפים אליהם.
Unofficial, LLM-ready knowledge base of the Grow Payments (Meshulam / grow.link) developer documentation — the complete Light API reference, OpenAPI schemas, integration guides and changelog, packaged as an
llms.txtcorpus and a Model Context Protocol (MCP) server.
Built for developers and AI agents (Claude, Cursor, etc.) that need accurate, grep-able answers about Grow / Meshulam payment integration without crawling a JavaScript documentation site every time. The original docs — and therefore much of this corpus — are written in a mix of English and Hebrew (עברית).
Naming: Grow and Meshulam (Hebrew גרואו / משולם) are the same payment provider — Meshulam rebranded to Grow. People search both; this repo covers the Grow / Meshulam Light API.
Why
The official docs at grow-il.readme.io are a client-rendered SPA — hard for LLMs and
scripts to consume, and mixed English/Hebrew. This repo turns the entire site into:
- 📄
data/llms.txt— a clean, grouped index of every page (the llms.txt standard). - 📚
data/llms-full.txt— the whole documentation set in one ~310 KB file. - 🗂️
data/pages/**.md— one Markdown file per page (guides + API reference). - 🔌
data/openapi/*.json— OpenAPI 3.0 schema for each API endpoint. - 🧩
data/openapi/_grow-payments-light-api.openapi.json+ Postman collection — the whole Light API (16 endpoints) merged into one importable spec. - 🤖
skills/grow-payments-mcp— an MCP server exposing it all as search tools.
What's inside
| | Count | |---|---| | Documentation pages | 100 | | Integration guides | 10 | | API reference pages | 84 | | API endpoints with OpenAPI schemas | 25 | | Changelog entries | 5 |
Covers: hosted payment pages, the Light Server REST API (regular & token payments, payment links, refunds, direct debit, transaction & payment-process info), webhooks / server-to-server callbacks, 3DS, NFC, POS devices, and the Apple / Bit SDKs.
.
├── data/ # the corpus (single source of truth)
│ ├── llms.txt # grouped index (start here)
│ ├── llms-full.txt # everything, one file
│ ├── index.json # machine-readable index
│ ├── pages/{home,docs,reference,changelog}/*.md
│ ├── openapi/*.json # per-endpoint OpenAPI 3.0 (+ _…light-api.openapi.json merged)
│ └── postman/*.json # Postman collection for the whole Light API
├── skills/
│ ├── grow-payments-docs/ # llms.txt skill (read files directly)
│ └── grow-payments-mcp/ # MCP server (search via tools)
└── tools/ # reproducible scraper (scrape.sh, extract.py, build_corpus.py)
Quick start
Option A — just read it (no install)
Point your LLM / agent / script at data/llms.txt and open the pages it
links to. For example:
git clone https://github.com/Shaharamir/grow-payments-mcp.git
cd grow-payments-mcp
grep -rIl -i "refund" data/pages # find pages
cat data/pages/reference/post_api-light-server-1-0-refundtransaction-1.md
Option B — MCP server (search tools for Claude / any MCP client)
Requires Python ≥ 3.10.
python3 -m venv .venv
.venv/bin/pip install -r skills/grow-payments-mcp/requirements.txt
.venv/bin/python skills/grow-payments-mcp/selftest.py # offline smoke test
Register with your MCP client (copy .mcp.json.example →
.mcp.json and set absolute paths):
{
"mcpServers": {
"grow-payments-docs": {
"command": "/abs/path/grow-payments-mcp/.venv/bin/python",
"args": ["/abs/path/grow-payments-mcp/skills/grow-payments-mcp/server.py"],
"env": { "GROW_DOCS_DATA": "/abs/path/grow-payments-mcp/data" }
}
}
}
Tools: search_docs, get_doc, list_docs, list_endpoints, get_openapi.
Resources: grow://llms.txt, grow://llms-full.txt, grow://openapi.json.
Option C — import the whole API
One merged OpenAPI 3.0 spec + a Postman collection for all 16 Light API endpoints:
- OpenAPI:
data/openapi/_grow-payments-light-api.openapi.json— import into Swagger UI, Insomnia, or any codegen. - Postman:
data/postman/grow-payments-light-api.postman_collection.json— Import → File in Postman.
Example prompt once connected: "Using grow-payments-docs, show the request fields for
createPaymentProcessand the refund endpoint's OpenAPI schema."
Grow API cheat-sheet
- Endpoints: server →
/api/light/server/1.0/...; Android/POS →/api/light/andriod/1.0/...(vendor spelling). - Sandbox host:
https://sandbox.meshulam.co.il(aliashttps://sandboxapi.grow.link). Production hosts are issued by Grow support. - Auth: per-business
userId+pageCode(and/or API key). Values shown in the docs are examples only. - Calls are server-to-server, usually
multipart/form-data; client-side requests are blocked.
Keeping it fresh
The corpus is regenerated from the live site (ReadMe.io embeds page content + OpenAPI in a
ssr-props JSON blob, which the scraper parses):
python3 -m pip install markdownify beautifulsoup4
./tools/scrape.sh # incremental; FORCE=1 ./tools/scrape.sh re-downloads all
⚠️ Disclaimer — this is NOT official
We are not Grow Payments / Meshulam. This is an independent, community-made, unofficial mirror. It is not affiliated with, authorized by, sponsored by, or endorsed by Grow Payments, Meshulam, or grow.link in any way.
All product names, trademarks, logos and brands — including the Grow logo shown above — are the property of their respective owners and are used here purely to identify the documentation this project mirrors (nominative use), not to imply any affiliation.
Documentation content under data/ is © Grow Payments and reproduced here for developer
convenience; the repository code is MIT-licensed (see LICENSE). Always verify
money- or security-critical details against the official docs
or Grow support before using in production. Rights holders: open an issue and any
content will be amended or removed promptly.
בעברית: זהו פרויקט לא רשמי ועצמאי. איננו Grow / משולם ואיננו מסונפים אליהם, מורשים או נתמכים על ידם. כל הסימנים המסחריים והלוגו שייכים לבעליהם. לבעלי הזכויות: פתחו issue ונסיר/נתקן כל תוכן לפי בקשה.
Snapshot: 2026-06-15.