Quickstart: Chat Memory

Build persistent chat memory using sessioned messages, compression, and memory search.

What You Will Build

Store chat messages by session
Retrieve compressed conversation context for LLM prompts
Search long-term memory with graph-aware retrieval

Prerequisites

PAPR_MEMORY_API_KEY configured in your environment
A stable sessionId and external_user_id per conversation

Minimal Setup

Store one chat message with process_messages=true.
Retrieve compressed context for the same session.
Run a memory search before response generation.

1) Store a Message

curl -X POST https://memory.papr.ai/v1/messages \
  -H "X-API-Key: $PAPR_MEMORY_API_KEY" \
  -H "Content-Type: application/json" \
  -H "X-Client-Type: curl" \
  -d '{
    "sessionId": "session_support_001",
    "role": "user",
    "content": "I prefer email notifications and weekly summaries.",
    "external_user_id": "user_123",
    "process_messages": true
  }'

2) Retrieve Session Context

curl -X GET "https://memory.papr.ai/v1/messages/sessions/session_support_001/compress" \
  -H "X-API-Key: $PAPR_MEMORY_API_KEY" \
  -H "X-Client-Type: curl"

Use context_for_llm directly in your model prompt.

3) Search Memory Before Responding

curl -X POST "https://memory.papr.ai/v1/memory/search?max_memories=20&max_nodes=15" \
  -H "X-API-Key: $PAPR_MEMORY_API_KEY" \
  -H "Content-Type: application/json" \
  -H "X-Client-Type: curl" \
  -d '{
    "query": "What communication preferences does this user have?",
    "external_user_id": "user_123",
    "enable_agentic_graph": true
  }'

Recommended Defaults

process_messages: true
enable_agentic_graph: true
max_memories: 15-20
max_nodes: 10-15

Validation Checklist

The stored message is visible in session history.
Compression endpoint returns session summary fields.
Search returns user preference context from prior turns.

Troubleshooting

If context is missing, confirm the same sessionId is used across writes and reads and review Messages Management.