Dietary restriction venue matcher
Match a user's dietary needs to venues that can actually accommodate them.
A user says 'I'm celiac and my partner is vegan — where can we eat in Canggu?' The naive approach searches for 'gluten-free vegan Canggu' and returns garbage. This prompt reads both users' dietary memory, queries the venue knowledge graph for menu-level dietary data, and returns only venues where BOTH diets are genuinely accommodated — not just 'they have a salad'.
The prompt distinguishes between venues that list dietary options on their menu (high confidence) and venues where the agent inferred dietary capability from reviews or descriptions (medium confidence). It never returns low-confidence matches without flagging them.
The prompt
<<<
You match users with dietary restrictions to venues that genuinely accommodate them.
INPUTS:
- user dietary restrictions (from user memory: diet, allergies)
- companion dietary restrictions (if provided)
- area (e.g. "Canggu", "Seminyak")
- meal type (breakfast | lunch | dinner | drinks)
SEQUENCE:
1. Read user memory for diet + allergies. If companion is an AGNT user, read theirs too.
2. Query venue knowledge graph for venues in the area with menu-level dietary tags.
3. For each candidate venue, check:
a. Does the menu explicitly list items matching EACH diner's restrictions?
b. Confidence: HIGH if menu data, MEDIUM if inferred from reviews, LOW if guessed.
4. Filter: only return venues where ALL diners have at least one HIGH or MEDIUM match.
5. Rank by: number of matching menu items * confidence weight.
OUTPUT:
{
"matches": [
{ "venue": "...", "confidence": "high", "matching_items": ["..."], "notes": "..." }
],
"no_match_reason": null | "no venues in area match all restrictions"
}
NEVER return a venue where you're guessing it might accommodate a restriction.
>>>When to use
Use this whenever a booking request includes dietary restrictions for one or more diners. It prevents the embarrassing 'we recommended a steakhouse to a vegan' failure.
Works with
Related guides
Configure your venue agent's memory and knowledge
Upload menus, set booking rules, and train a custom personality so your venue agent answers like your best staff member.
How to use Claude Code with AGNT
Install Claude Code, register AGNT as an MCP server, and start calling real venue tools from inside your terminal.
MCP integration guide
Connect any MCP-capable host — Claude Code, Cursor, Zed, or your own client — to AGNT's live tools.
Share as social post
Dietary restriction venue matcher (mcp + claude-code) — Match a user's dietary needs to venues that can actually accommodate them. https://agntdot.com/prompts/dietary-restriction-matcher
187 / 280 chars
Related prompts
MCP memory-write template
Structured prompt for writing durable facts into AGNT user memory via MCP.
MCP graph-query template
Query AGNT's knowledge graph through MCP with clean fallbacks.
RAG recall guard prompt
Prevent semantic recall from injecting stale or contradictory facts into the agent's prompt.