API Endpoints

De KB Search Agent communiceert met de WATCH API V2 op aipromptpower.com. Hieronder staan alle endpoints die de agent gebruikt.

Authenticatie

POST /api/v2/auth/login
Content-Type: application/json

{
    "employee_code": "REINT",
    "password": "..."
}

Response: { "data": { "access_token": "eyJ..." } }

Chat (semantisch zoeken + LLM antwoord)

POST /api/v2/chat/WATCH-KB-SEARCH/{template_code}
Authorization: Bearer {token}
Content-Type: application/json

{
    "question": "Hoe werkt de artikelbijlage?",
    "conversation_id": "kb-agent-2026-03-27-11-00"
}

De conversation_id koppelt vervolgvragen aan hetzelfde gesprek. Wordt gelogd in CHAT_CONVERSATIONS.

Keyword Search (bronnen zoeken op tekst)

GET /api/v2/knowledge-bases/{kb_code}/sources?search={term}&status=L&limit=20
Authorization: Bearer {token}

Doorzoek elke KB apart. Combineert resultaten uit meerdere KB's.

Volledige Brontekst

GET /api/v2/knowledge-bases/{kb_code}/sources/{source_code}
Authorization: Bearer {token}

Retourneert complete_content (LONGTEXT) met de volledige brontekst.

Chunk Search

GET /api/v2/knowledge-bases/{kb_code}/sources/{source_code}/chunks?search={term}&limit=20
Authorization: Bearer {token}

Context Blocks (beheer)

GET    /api/v2/context-blocks
GET    /api/v2/context-blocks/{code}
POST   /api/v2/context-blocks
PATCH  /api/v2/context-blocks/{code}
DELETE /api/v2/context-blocks/{code}

AI Models (beheer)

GET    /api/v2/ai-models
GET    /api/v2/ai-models/{code}
POST   /api/v2/ai-models
PATCH  /api/v2/ai-models/{code}
POST   /api/v2/ai-models/{code}/costs

GET    /api/v2/ai-model-settings
GET    /api/v2/ai-model-settings/{code}
POST   /api/v2/ai-model-settings
PATCH  /api/v2/ai-model-settings/{code}

Base URL

Alle endpoints zijn relatief ten opzichte van:

https://www.aipromptpower.com/source/api-v2/public/api/v2/