Tapsi Food

For AI agents

• What: unofficial MCP server (model context protocol) for Tapsi Food (tapsi.food), food and grocery delivery in Iran. Not Snappfood. Queries in Persian match best; prices in Toman.
• Install: uvx tapsifood-mcp or claude mcp add tapsifood -- uvx tapsifood-mcp. Transport: stdio. Needs uv (Python 3.10+).
• Auth: none, no API key or account. Tapsi Food hangs connections from non-Iranian IPs: run it from Iran without a VPN, or set TAPSIFOOD_MCP_PROXY to an HTTP proxy that exits in Iran.
• Access: read-only. No login, basket, orders or payment, so no confirmation is needed before calling any tool.
• Tools: 11, all prefixed tf_. Most need lat and long (not tf_reviews, tf_find_location, tf_list_cities). Typical sequence: tf_find_location -> tf_search_dishes -> tf_vendor_info (-> tf_menu, tf_reviews).
• Use when: the user wants to find a dish or restaurant near an address in Tehran, Mashhad or Shiraz, compare prices and fees, read a menu or reviews, or find discounts.
• Do not use when: the user wants to order, pay or log in, or asks about a city Tapsi Food does not serve.
• Coverage: Tehran, Mashhad, Shiraz. Elsewhere the API answers 404 or OutOfCoverage.
• Output: compact JSON. Prices in Toman, ratings 0-5. Null and empty fields are omitted (a missing rating = not rated). Menu descriptions are cut to 100 characters and review comments to 250 (full_text: true for the whole text). Replies stay under about 15 KB at the maximum limit.
• Machine-readable: server.json and the iran-mcp catalog (TOOLS.md, servers.json, llms.txt).

Why

Picking dinner on a delivery app means opening a restaurant, scrolling the menu, then finding out the delivery fee and minimum order. Asking "what is the cheapest burger near me that is open now, and what will it really cost?" takes many taps. With tapsifood-mcp an agent finds the dish, then checks the restaurant's fee, packaging and minimum order for your address, and answers with the true cost: dish price minus discount plus packaging plus delivery fee.

You: Cheapest pizza near Valiasr Square, and is it worth the delivery fee?

Agent: calls tf_find_location -> tf_search_dishes(sort="cheapest") -> tf_vendor_info

It lists the cheapest open matches with price and discount, then checks the chosen restaurant's delivery fee to your point, its minimum order and its status.

What it can do

• Geocode a place name or address to coordinates, reverse a point back to an address, and list the cities Tapsi Food shows
• Search dishes by name near a delivery point, sorted by relevance, cheapest, biggest discount, rating or distance
• Search restaurants and stores by name or cuisine near a point
• Browse restaurants or grocery stores with rating, delivery fee, minimum order and ETA, filtered by cuisine, discount, courier or new vendors
• Catch deals: Tap Time (تپ تایم) flash deals and the Food Shots (فودشات) discount campaign
• Read a live menu: every dish with price, discount and stock
• Check a vendor before ordering: delivery fee to your point, minimum order, packaging, ETA, open or closed, hours, address and rating spread
• Read reviews with the rating of each ordered dish
• Read-only by design: no login, no basket, no orders, no payment

Example prompts:

• "Find the cheapest burger near Enghelab Square in Tehran that is open now."
• "Which restaurants near me have a discount and a delivery fee under 30,000 Toman?"
• "Show the menu of this restaurant and tell me its minimum order."
• "What are the Tap Time flash deals around Azadi Tower right now?"
• "Read the latest reviews of this restaurant. Are people happy with the delivery?"
• ارزان‌ترین پیتزا نزدیک میدان ولیعصر کدام است؟
• رستوران‌های تخفیف‌دار نزدیک من را با هزینه ارسال و حداقل سفارش نشان بده.
• تپ تایم امروز چه غذاهایی دارد؟

Quick start

You need uv. No API key or account. Run it from Iran, or add TAPSIFOOD_MCP_PROXY (see Configuration).

claude mcp add tapsifood -- uvx tapsifood-mcp# behind a proxy that exits in Iran:claude mcp add tapsifood -e TAPSIFOOD_MCP_PROXY=http://127.0.0.1:8080 -- uvx tapsifood-mcp

{  "mcpServers": {    "tapsifood": { "command": "uvx", "args": ["tapsifood-mcp"] }  }}

{  "servers": {    "tapsifood": { "type": "stdio", "command": "uvx", "args": ["tapsifood-mcp"] }  }}

codex mcp add tapsifood -- uvx tapsifood-mcp

[mcp_servers.tapsifood]command = "uvx"args = ["tapsifood-mcp"]

{  "mcpServers": {    "tapsifood": { "command": "uvx", "args": ["tapsifood-mcp"] }  }}

Tools

Start with a place: tf_find_location or tf_list_cities give the lat and long the other tools need. Vendor codes look like 2y97wk.

All 11 tools are annotated readOnlyHint: true and return compact JSON.

How it works

  AI agent  (Claude, Cursor, Copilot, ...)      │      │  MCP over stdio      ▼  tapsifood-mcp  (runs on your machine)      │      │  HTTPS      └──────▶  api.tapsi.food   dishes, vendors, menus, reviews, deals

tapsifood-mcp runs locally and calls the same endpoints the tapsi.food web app uses. There is no hosted server in between and no API key. Requests go out through a Chrome-impersonating client, at most 2 at a time with a little jitter. GET answers are cached for 60 seconds, a timeout, network error, 5xx or 429 is retried up to twice, and it only talks to api.tapsi.food. It only reads.

Good to know

• Prices are in Toman, as the API sends them. Ratings are 0-5.
• Delivery depends on the point. Menus, fees and coverage change with the latitude and longitude you pass. HTTP 200 does not mean the vendor delivers there: delivery_fee: null means out of coverage.
• Coverage: Tehran, Mashhad and Shiraz. Elsewhere feeds answer 404, search finds nothing and the menu says OutOfCoverage.
• Network: Tapsi Food hangs connections from non-Iranian IPs (a timeout on every call is the sign). Run it from Iran without a VPN, or set TAPSIFOOD_MCP_PROXY. System proxy variables are ignored on purpose.
• Open vendors only in search. Dish and vendor search list only restaurants open now (the site does the same), so at night total can be 0. tf_restaurants and tf_discounted_vendors also list closed vendors, flagged open: false.
• Search-index prices can lag the menu: confirm with tf_menu. Menu and profile minimums sometimes differ; tf_vendor_info returns both (min_order, min_order_profile).
• Sorting dishes: top_rated, cheapest and biggest_discount sort the first 100 matches, not the whole result set.
• ETA is an upper bound: eta_max is the site's "up to N minutes"; the API sends no lower bound.
• Reviews: total counts every review including rating-only ones; only reviews with a comment are listed, so a page can hold fewer than limit (keep paging while has_more). Writer names are left out on purpose.
• Browse pages: page 1 also carries filter and banner widgets, so it lists fewer vendors than page_size.
• What it does not do: no login, basket, orders or payment. Coupons are not exposed by the site's anonymous API.

FAQ

npx @modelcontextprotocol/inspector uvx tapsifood-mcp

Configuration

فارسی

claude mcp add tapsifood -- uvx tapsifood-mcp

Development

git clone https://github.com/sepehr071/tapsifood-mcp && cd tapsifood-mcpuv syncuv run ruff check . && uv run ruff format --check .uv run python -m pytest -q          # offline unit testsuv run python -m pytest -m live -q  # real API; needs an Iranian IPuv build

Tools live in src/tapsifood_mcp/location.py, discover.py and vendor.py; each is a typed async function with a docstring that tells the agent when to use it. Issues and PRs are welcome.

Releases: bump the version in pyproject.toml and server.json, then push a v* tag. GitHub Actions tests, publishes to PyPI and the MCP Registry, and creates the GitHub Release.

Related servers

• iran-mcp: catalog of all Iranian-service MCP servers
• Snapp MCP server for Snappfood and SnappMarket food and grocery delivery
• Okala MCP server for online grocery shopping
• Digikala MCP server for Iran's largest online store
• Torob MCP server for price comparison across Iranian shops
• Listed on mcpfarsi.ir, the Persian MCP marketplace

Disclaimer

Unofficial and not affiliated with or endorsed by Tapsi. It is read-only and uses the endpoints of the tapsi.food web app, which can change without notice. Please keep request rates reasonable.

Author

[Sepehr Radmard: open-source MCP servers for the Iranian web. LinkedIn: linkedin.com/in/sepehr-radmard]

Built and maintained by Sepehr Radmard (سپهر رادمرد). Connect on LinkedIn or follow @sepehr071 on GitHub for more MCP servers for Iranian websites, all listed in iran-mcp. If this saves you time, a star on the repo helps other people find it.

License

MIT