Get context / SLM prompt (Step 4)

curl -X POST https://api.forboc.ai/npcs/npcId/context \
     -H "Authorization: Bearer <token>" \
     -H "Content-Type: application/json" \
     -d '{
  "memories": [
    {},
    {},
    {}
  ],
  "observation": "The NPC notices a suspicious figure lurking near the village entrance.",
  "npcState": {},
  "persona": "A cautious village guard who prioritizes the safety of the townsfolk and is wary of strangers."
}'

{
  "prompt": "You are a cautious village guard. You have noticed a suspicious figure lurking near the village entrance. Recall your memories: the recent thefts in the village, your training in identifying threats, and your duty to protect the villagers. Decide your next action carefully.",
  "constraints": {
    "maxTokens": 150,
    "temperature": 0.7,
    "stop": [
      "\n"
    ]
  },
  "ctxMessages": [
    {
      "pmRole": "system",
      "pmContent": "You are a cautious village guard tasked with protecting the village."
    },
    {
      "pmRole": "assistant",
      "pmContent": "What do you do upon seeing the suspicious figure?"
    }
  ]
}

SDK sends recalled memories from its local vector DB. API synthesizes a full SLM prompt from persona + rules + memories + observation, with generation constraints.

Authentication

AuthorizationBearer

Bearer authentication of the form Bearer <token>, where token is your auth token.

Path parameters

npcIdstringRequired

Request

This endpoint expects an object.

memorieslist of objectsRequired

Memories recalled by the SDK from its local vector DB

observationstringRequired

Original observation (echoed from directive step)

npcStatemap from strings to anyRequired

Current NPC state

personastringRequired

NPC persona used to construct context

rulesetIdstring or nullOptional

Response

Full SLM prompt and constraints

promptstring

Full SLM prompt text — SDK runs this verbatim

constraintsobject

ctxMessageslist of objects or null