373,000+
US restaurants
Focused restaurant coverage for launch.
Compass DaaS
Vegan suitability needs evidence, not tags.
Compass helps AI agents search US restaurants, enrich place records, and make conservative vegan and plant-based fit decisions with structured evidence, confidence, and reason codes.
The explainable vegan and plant-based decision layer for AI agents. Built for food, travel, hospitality, local discovery, and agent products.
POST /v1/searchCompact proof
{
"match_for_profile": {
"decision": "fit",
"confidence": "high"
},
"reason_codes": [
"VEGAN_MENU_FULL_RESTAURANT",
"CROSS_CONTAMINATION_UNKNOWN"
],
"risk_flags": ["limited_evidence"]
}US-only
Current geography
Clear scope for product and buyer trust.
Vegan-first
Plant-based depth
Strongest coverage is vegan suitability.
3 APIs
Search, Enrich, Decide
Use one surface or combine the workflow.
POST /v1/searchCompact proof
{
"match_for_profile": {
"decision": "fit",
"confidence": "high"
},
"reason_codes": [
"VEGAN_MENU_FULL_RESTAURANT",
"CROSS_CONTAMINATION_UNKNOWN"
],
"risk_flags": ["limited_evidence"]
}Structured developer surfaces
Choose the surface that matches your workflow.
Search, Enrich, and Decision are the primary REST APIs. Compass MCP is the agent-native wrapper for Claude, Cursor, Codex, and other MCP-compatible tools. Restaurant Details remains the supporting lookup when a Search or Enrich result needs the full structured record behind it.
Search API
POST /v1/search
107 structured fields
Use this when your agent needs restaurant candidates from a natural-language vegan or plant-based query.
Enrich API
POST /v1/enrich/restaurant
90 structured fields
Use this when you already have a restaurant or place record and want Compass context added to it.
Decision API
POST /v1/decision/restaurant-fit
27 decision fields
Use this when your product needs a conservative fit, not_fit, or unknown verdict for a user profile.
Compass MCP Server
npx -y @compass-food/mcp
3 agent tools
Use this when your agent workflow runs inside Claude Desktop, Cursor, Codex, or another MCP-compatible client.
Restaurant Details Lookup
GET /v1/restaurants/:id
176 structured fields
Use this after Search or Enrich when you need the full structured record behind a Compass result.
Field counts describe current v1 response depth by endpoint. They are not a guarantee that every restaurant has every field populated.
Why Compass
Built for decisions, not just place search.
Beyond flat tags
Compass separates weak category labels from stronger vegan and plant-based evidence, so your agent does not treat every vegan-friendly tag as equal.
Conservative by design
When evidence is stale, incomplete, or conflicting, Compass can return unknown instead of forcing false confidence.
Inspectable outputs
Reason codes, confidence, source freshness, and verification flags help developers debug and help agents explain their answers.
Concrete API examples
Full request and response shapes, after the trust context.
The examples are representative static payloads. Field availability varies by restaurant, and uncertainty is returned when evidence is missing.
Search requestPOST /v1/searchExample data
curl -X POST "https://api.compassfoodtechnologies.com/v1/search" \
-H "Content-Type: application/json" \
-H "X-Compass-API-Key: $COMPASS_API_KEY" \
-H "X-Compass-Mode: rich" \
-d '{
"query": "strict vegan bowls near Brooklyn Heights",
"location": {
"lat": 40.6959,
"lng": -73.9936,
"radius_m": 3000
},
"user_profile": {
"diet": "strict_vegan",
"exclude_cross_contamination": false
},
"limit": 3,
"include_evidence": true
}'Search responseJSONExample data
{
"compass_request_id": "req_example_green_bowl_search",
"query_interpretation": {
"diet_inferred": "strict_vegan",
"cuisine_inferred": "bowls",
"dish_inferred": "grain bowl",
"intent_inferred": "dish",
"target_inferred": "grain bowl",
"location_inferred": "40.6959,-73.9936",
"radius_m": 3000,
"radius_m_inferred": false
},
"results": [
{
"compass_id": "cmp_example_green_bowl_kitchen_001",
"name": "Green Bowl Kitchen",
"address": {
"street": "123 Main Street",
"city": "Brooklyn",
"state": "NY",
"postal_code": "11201",
"country": "US"
},
"coordinates": {
"lat": 40.6957,
"lng": -73.9934
},
"distance_m": 38,
"vegan_score": {
"overall": 91,
"scoring_version": "v3.0-deterministic",
"dimensions": {
"D1_vegan_coverage": {
"score": 96,
"weight": 0.25,
"confidence": "high",
"evidence_ids": [
"ev_cmp_example_green_bowl_menu"
]
},
"D4_cross_contamination": {
"score": 40,
"weight": 0.2,
"confidence": "low",
"evidence_ids": [
"ev_cmp_example_green_bowl_cross_contamination_unknown"
]
}
}
},
"dietary_profile": {
"fully_vegan": true,
"vegan_friendly": true
},
"cuisine_types": ["plant-based", "bowls", "salads"],
"last_verified_at": "2026-04-29T12:00:00Z",
"last_evaluated_at": "2026-04-29T12:00:00Z",
"source_freshness": {
"oldest_signal_age_days": 14,
"newest_signal_age_days": 14,
"recommended_refresh": false
},
"evidence": [
{
"evidence_id": "ev_cmp_example_green_bowl_menu",
"type": "menu_item",
"source": "compass://deterministic-evidence/menu_dishes",
"weight": "primary",
"tier": "A",
"excerpt": "Menu section lists fully vegan bowls, sauces, and sides."
},
{
"evidence_id": "ev_cmp_example_green_bowl_cross_contamination_unknown",
"type": "review_signal",
"source": "compass://deterministic-evidence/review_signals",
"weight": "supporting",
"tier": "A",
"excerpt": "Cross-contamination handling not declared. Review signals do not mention shared equipment or dedicated prep area."
}
],
"match_for_profile": {
"decision": "fit",
"confidence": "high",
"reason_codes": [
"VEGAN_MENU_FULL_RESTAURANT",
"CROSS_CONTAMINATION_UNKNOWN",
"HIGH_CONFIDENCE_DECISION"
],
"risk_flags": ["limited_evidence"],
"verification_required": true,
"recommended_user_text": "Compass has strong menu evidence for a strict-vegan match. Cross-contamination handling is not declared in available evidence, so verify with the restaurant when that matters."
}
}
],
"total_results": 3,
"ai_generated": true,
"human_review_recommended": false,
"metadata": {
"mode": "rich",
"result_source": "qdrant_restaurants_v2"
}
}Enrich requestPOST /v1/enrich/restaurantExample data
curl -X POST "https://api.compassfoodtechnologies.com/v1/enrich/restaurant" \
-H "Content-Type: application/json" \
-H "X-Compass-API-Key: $COMPASS_API_KEY" \
-d '{
"name": "Green Bowl Kitchen",
"address": "123 Main Street, Brooklyn, NY 11201"
}'Enrich responseJSONExample data
{
"compass_request_id": "req_example_green_bowl_enrich",
"matched": true,
"match_confidence": 0.91,
"compass_id": "cmp_example_green_bowl_kitchen_001",
"restaurant": {
"compass_id": "cmp_example_green_bowl_kitchen_001",
"name": "Green Bowl Kitchen",
"address": {
"street": "123 Main Street",
"city": "Brooklyn",
"state": "NY",
"postal_code": "11201",
"country": "US"
},
"coordinates": {
"lat": 40.6957,
"lng": -73.9934
},
"vegan_score": {
"overall": 91,
"scoring_version": "v3.0-deterministic",
"dimensions": {
"D1_vegan_coverage": {
"score": 96,
"weight": 0.25,
"confidence": "high",
"evidence_ids": [
"ev_cmp_example_green_bowl_menu"
]
},
"D6_restaurant_quality": {
"score": 90,
"weight": 0.1,
"confidence": "high",
"evidence_ids": [
"ev_cmp_example_green_bowl_quality"
]
}
}
},
"dietary_profile": {
"fully_vegan": true,
"vegan_friendly": true
},
"cuisine_types": ["plant-based", "bowls", "salads"],
"last_verified_at": "2026-04-29T12:00:00Z",
"last_evaluated_at": "2026-04-29T12:00:00Z",
"source_freshness": {
"oldest_signal_age_days": 14,
"newest_signal_age_days": 14,
"recommended_refresh": false
},
"evidence": [
{
"evidence_id": "ev_cmp_example_green_bowl_menu",
"type": "menu_item",
"source": "compass://deterministic-evidence/menu_dishes",
"weight": "primary",
"tier": "A"
},
{
"evidence_id": "ev_cmp_example_green_bowl_quality",
"type": "review_signal",
"source": "compass://deterministic-evidence/review_signals",
"weight": "supporting",
"tier": "A",
"excerpt": "Recent customer signals mention consistent vegan menu availability."
}
]
},
"evidence": [
{
"evidence_id": "ev_cmp_example_green_bowl_menu",
"type": "menu_item",
"source": "compass://deterministic-evidence/menu_dishes",
"weight": "primary",
"tier": "A"
}
],
"ai_generated": true
}Decision requestPOST /v1/decision/restaurant-fitExample data
curl -X POST "https://api.compassfoodtechnologies.com/v1/decision/restaurant-fit" \
-H "Content-Type: application/json" \
-H "X-Compass-API-Key: $COMPASS_API_KEY" \
-H "X-Compass-Mode: rich" \
-d '{
"compass_id": "cmp_example_green_bowl_kitchen_001",
"user_profile": {
"diet": "strict_vegan",
"exclude_cross_contamination": false
}
}'Decision responseJSONExample data
{
"compass_request_id": "req_example_green_bowl_decision",
"compass_id": "cmp_example_green_bowl_kitchen_001",
"last_evaluated_at": "2026-04-29T12:00:00Z",
"decision": "fit",
"confidence": "high",
"reason_codes": [
"VEGAN_MENU_FULL_RESTAURANT",
"SOURCE_CONFIDENCE_HIGH",
"CROSS_CONTAMINATION_UNKNOWN",
"HIGH_CONFIDENCE_DECISION"
],
"risk_flags": ["limited_evidence"],
"evidence": [
{
"evidence_id": "ev_cmp_example_green_bowl_menu",
"type": "menu_item",
"source": "compass://deterministic-evidence/menu_dishes",
"weight": "primary",
"tier": "A"
},
{
"evidence_id": "ev_cmp_example_green_bowl_cross_contamination_unknown",
"type": "review_signal",
"source": "compass://deterministic-evidence/review_signals",
"weight": "supporting",
"tier": "A"
}
],
"source_freshness": {
"oldest_signal_age_days": 14,
"newest_signal_age_days": 14,
"recommended_refresh": false
},
"recommended_user_text": "Compass found strong vegan menu evidence. Cross-contamination handling is not declared in available evidence, so verify details with the restaurant when constraints are critical.",
"verification_required": true,
"human_review_recommended": false,
"ai_generated": true
}Where Compass fits
Pick the path that matches your product.
Start with one surface, then add deeper context as your workflow moves from discovery to explanation and user-specific decisions.
I am building a restaurant search or travel agent.
Start with Search API.
I already have restaurant records from another provider.
Use Enrich API to add vegan and plant-based intelligence.
I need to decide if a place fits a user.
Use Decision API.
I am building with Claude, Cursor, or other agent tools.
Use the Compass MCP server.
Responsible plant-based output
Clear uncertainty is part of the product.
Compass is not an allergy-certification service and does not replace restaurant confirmation for safety-sensitive decisions. It is designed to expose evidence, uncertainty, and conservative risk signals so your product can handle vegan and plant-based questions responsibly.
Build your first plant-based restaurant answer today.
Create a free Sandbox account, make a Search request, then add Enrich or Decision when your workflow needs deeper context.