Quick Start

1. Install PRISM

pip install prism-platform

2. Authenticate

prism login
# Opens browser → enter the code → done

Or set an API key directly:

export MARC27_API_KEY=m27_your_key_here

3. Search the Knowledge Graph

prism research "Find materials with high creep resistance" --depth 0

--depth 0 searches the local graph only (free). Higher depths enable web search (costs LLM calls).

4. Try GraphQL

curl -X POST https://api.marc27.com/api/v1/graphql \
  -H "X-API-Key: m27_your_key" \
  -H "Content-Type: application/json" \
  -d '{"query":"{ search(term: \"titanium\", limit: 5) { name entityType label } }"}'

Response:

{
  "data": {
    "search": [
      {"name": "Titanium", "entityType": "CHM", "label": "Chemical"},
      {"name": "α-Titanium", "entityType": "CHM", "label": "Chemical"},
      {"name": "β-Titanium", "entityType": "CHM", "label": "Chemical"}
    ]
  }
}

5. Explore Entity Relationships

curl -X POST https://api.marc27.com/api/v1/graphql \
  -H "X-API-Key: m27_your_key" \
  -H "Content-Type: application/json" \
  -d '{"query":"{ entity(name: \"Nickel\") { name neighbors(limit: 3) { target { name entityType } relType count } } }"}'

This returns Nickel’s neighbors — what it co-occurs with across 5 million papers:

{
  "data": {
    "entity": {
      "name": "Nickel",
      "neighbors": [
        {"target": {"name": "Ni", "entityType": "CHM"}, "relType": "CHM-CHM", "count": 14664},
        {"target": {"name": "SEM", "entityType": "CMT"}, "relType": "CHM-CMT", "count": 6972},
        {"target": {"name": "Catalyst", "entityType": "APL"}, "relType": "CHM-APL", "count": 6571}
      ]
    }
  }
}

Next Steps

• GraphQL Reference — all 20 queries and 6 mutations
• Research Mode — how the RLM engine works
• Model Deployments — deploy models from HuggingFace