API Reference
Complete reference for all API endpoints.
Base URL: https://api.lthn.ai/v1
Brain Memory
Store and retrieve agent memories with vector search. Powered by Qdrant for semantic retrieval.
/brain/remember
Store a new memory in the vector database.
Request Body
{"content": "Go uses structural typing", "type": "fact", "project": "go-agentic", "tags": ["go", "typing"]}
Response
{"id": "mem-abc-123", "type": "fact", "project": "go-agentic", "created_at": "2026-03-03T12:00:00+00:00"}
/brain/recall
Search memories by semantic query. Returns ranked results with confidence scores.
Request Body
{"query": "how does typing work in Go", "top_k": 5, "project": "go-agentic"}
Response
{"memories": [{"id": "mem-abc-123", "type": "fact", "content": "Go uses structural typing", "confidence": 0.95}], "scores": {"mem-abc-123": 0.87}}
/brain/forget/{id}
Delete a specific memory by ID.
Response
{"deleted": true}
Content Scoring
Score text for AI patterns and analyse linguistic imprints via the EaaS scoring engine.
/score/content
Score text for AI-generated content patterns. Returns a score (0-1), confidence, and label.
Request Body
{"text": "The text to analyse for AI patterns", "prompt": "Optional scoring prompt"}
Response
{"score": 0.23, "confidence": 0.91, "label": "human"}
/score/imprint
Perform linguistic imprint analysis on text. Returns a unique imprint fingerprint.
Request Body
{"text": "The text to analyse for linguistic patterns"}
Response
{"imprint": "abc123def456", "confidence": 0.88}
Collections
Manage vector database collections for brain memory storage.
/brain/collections
Ensure the workspace collection exists in the vector database. Creates it if missing.
Response
{"status": "ok"}
Health
Health check endpoints. These do not require authentication.
/score/health
Check the health of the scoring engine and its upstream services.
Response
{"status": "healthy", "upstream_status": 200}