Persistent memory layer with entity resolution, temporal features, and graph-aware recall. Self-host in 5 minutes. No vendor lock-in.
from remembra import Remembra
# Initialize client
client = Remembra(api_key="your-key")
# Store a memory with entity extraction
client.store(
content="John met Sarah at the AI conference in NYC",
user_id="user-123"
)
# Recall with semantic + entity search
memories = client.recall(
query="Who did John meet?",
user_id="user-123"
)
# → Returns: John met Sarah at the AI conferenceFeatures that actually matter for production AI
Automatically extract and link entities. "Mr. Kim" and "David Kim" resolve to the same person.
Memories that understand time. TTL, natural decay, and historical queries built-in.
Not just vectors. Combine semantic search with BM25 and graph traversal for better recall.
Built with security in mind. API keys, rate limiting, user scoping, and audit logs.
One-command Docker deployment. Your data stays on your infrastructure.
Visual memory browser with entity graphs, search, and real-time monitoring.
Remembra vs. the alternatives
| Feature | Remembra | Mem0 | Zep |
|---|---|---|---|
| Entity Resolution | ✓ Built-in | Partial | ✗ |
| Temporal Features (TTL, Decay) | ✓ Full | ✗ | TTL only |
| Hybrid Search (BM25 + Vector) | ✓ | ✗ | ✓ |
| Graph-Aware Recall | ✓ | Graph Memory | ✗ |
| Self-Host (Full Features) | ✓ 100% | Limited | ✓ |
| Dashboard UI | ✓ Included | Pro only | ✓ |
| Open Source | ✓ Apache 2.0 | ✓ | ✓ |
Build AI that actually remembers
Give your chatbots persistent memory across sessions. Remember user preferences, past conversations, and context.
Patient history, treatment plans, and medical context that persists and connects. HIPAA-ready with full audit logs.
Adaptive learning that remembers student progress, struggles, and breakthroughs. Personalized tutoring at scale.
Shopping assistants that remember preferences, past purchases, and browsing behavior for personalized recommendations.
AI that knows your leads, their history, and relationship context. Turn cold outreach into warm conversations.
Long-running agents that maintain state, learn from actions, and build knowledge over time.
Start free. Scale as you grow.
Self-host on your own infrastructure
Managed hosting, zero ops
For large-scale deployments
Deploy in 5 minutes. Open source. No vendor lock-in.