Ranks animals by suffering per individual (welfare range × P(sentience) × condition intensity × duration); also compares foods per kg, kcal or protein. Suffering Footprint is built for agents as well as people. Ask an agent to estimate a diet, compare substitutes, or explain why one food has a larger footprint than another, then follow the sources.
Remote MCP server
Endpoint: https://sufferingfootprint.com/mcp. Transport: Streamable HTTP, JSON responses. It is stateless, read-only and needs no key. It supports MCP revision 2026-07-28 and the earlier initialize-based revisions back to 2024-11-05.1
Config
{
"mcpServers": {
"suffering-footprint": {
"url": "https://sufferingfootprint.com/mcp"
}
}
}
If your client supports only local stdio servers, use its remote-server bridge if it has one, or call the JSON API directly.
Tools
| Tool | Use it for |
|---|---|
rank_animals_per_individual | Which animals suffer most, per animal: score with range, per-day intensity, band, estimates (assume) and sources. Options: include (food, lab, all), vertebrates_only, include_slaughter. |
calculate_suffering_footprint | A meal, recipe, shopping list or diet. Items take grams, kcal or protein_g; times scales it, for example 365 for a year. |
compare_foods | Rank foods per animal (basis: "individual"), per kg, 1,000 kcal or 100 g protein. Per-kg bases count animals per unit of food. |
list_foods | All foods, aliases, tiers and footprints. |
get_species_info | Every input, estimate and source for one food or animal, with its per-animal score. |
explain_method | Painparity, both formulas, units and exclusions. |
Test it with curl
curl -s https://sufferingfootprint.com/mcp \
-H 'Content-Type: application/json' \
-H 'Accept: application/json, text/event-stream' \
-H 'MCP-Protocol-Version: 2026-07-28' \
-H 'Mcp-Method: tools/call' -H 'Mcp-Name: compare_foods' \
-d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"compare_foods","arguments":{"foods":["beef","chicken","shrimp"],"basis":"protein"},"_meta":{"io.modelcontextprotocol/protocolVersion":"2026-07-28"}}}'
JSON API
| Request | Returns |
|---|---|
GET /api/hierarchy | Animals ranked by suffering per individual; include=food|lab|all, vertebrates_only, include_slaughter |
GET /api/foods | All foods with footprints and aliases |
GET /api/foods/chicken | All inputs and sources for one food |
GET /api/footprint?food=chicken&grams=200 | One food; also kcal=, protein_g= or items=chicken:200,rice:150 |
POST /api/footprint | A meal or diet: {"items":[{"food":"eggs","grams":120}],"times":365} |
GET /api/compare?foods=beef,chicken,shrimp&basis=kcal | A ranking; basis individual, kg, kcal (default) or protein |
GET /api/method | Formula, units and exclusions |
GET /api/data | The complete dataset with sources |
curl -s -X POST https://sufferingfootprint.com/api/footprint \
-H 'Content-Type: application/json' \
-d '{"items":[{"food":"chicken","grams":150},{"food":"eggs","grams":100},{"food":"tofu","grams":200}],"times":365}'
Machine-readable docs: OpenAPI and llms.txt. CORS is enabled.
Guidance for agents
- For “which animals suffer most”, use the per-individual ranking. Per-kg figures count animals per kilogram.
- Say when a value is an editorial estimate (listed in
assume). - Report the plausible range as well as the central value.
- Say that the footprint covers direct suffering of farmed animals, not every effect of food production.
- Use ratios and rankings rather than presenting suffering-days as an exact measurement.
- Link users to the relevant food or method page when they want detail.
Sources
- Model Context Protocol specification, revision 2026-07-28: Streamable HTTP transport. modelcontextprotocol.io/specification/2026-07-28/basic/transports