Simple Pricing

Start free. Scale when you are ready.

No per query charges. No hidden fees. Every plan includes the full 5 layer hybrid search engine and automatic knowledge graph extraction.

Free

$0/ forever

Perfect for trying Hebbrix, building a proof of concept, or running a personal project.

1,000 memories
5 collections
5 layer hybrid search
Knowledge graph (auto)
OpenAI compatible API
3 tier memory architecture
Community support
Get Started Free
Most Popular

Pro

$49/ per month

For teams building production AI agents that need reliable memory at scale.

100,000 memories
Unlimited collections
5 layer hybrid search
Knowledge graph (auto)
OpenAI compatible API
3 tier memory architecture
Smart ingestion pipeline (infer:true)
Webhook notifications
Priority support
Usage analytics dashboard
Start Pro Trial

Enterprise

Custom/ contact us

For organizations that need dedicated infrastructure, custom SLAs, and hands on support.

Unlimited memories
Unlimited collections
Everything in Pro
Dedicated infrastructure
Custom SLA (99.99% uptime)
SSO and SAML authentication
Data residency options
Dedicated account manager
Custom integrations
On premise deployment option
Talk to Sales

Included in every plan

5 layer hybrid search

Semantic vectors, BM25 keywords, knowledge graph, importance scoring, and recency boosting. All in under 50ms.

3 tier memory

Short term, medium term, and long term memory with automatic promotion and natural decay.

Multi tenant isolation

Organize memories by user, application, or namespace. Full data isolation between tenants.

Encryption at rest

All memories encrypted with AES 256. Your data stays private and secure.

OpenAI compatible

Drop in replacement for OpenAI's API format. Change the base URL and you are done.

Knowledge graph

Automatic entity and relationship extraction. No configuration needed.

Common questions

What counts as a memory?

A memory is a single stored fact or piece of information. When you use infer:true mode, one conversation turn might produce multiple memories since Hebbrix extracts individual facts. For example, 'I live in Austin and love spicy food' becomes two memories.

Can I upgrade or downgrade at any time?

Yes. You can upgrade instantly and your new limits take effect right away. Downgrading happens at the end of your billing cycle. Your existing memories are never deleted when you downgrade.

Is there a per query charge?

No. Search queries are unlimited on all plans. You only pay based on how many memories you store, not how often you search them.

Do you offer a startup program?

Yes. If you are an early stage startup (pre Series A), reach out to us at hello@hebbrix.com. We offer extended free tiers and discounted Pro plans for startups building with AI memory.

What happens when I hit my memory limit?

You will get a warning at 80% capacity. Once you reach your limit, new memory creation is paused until you upgrade or free up space. Existing memories and search continue working normally.

How does Hebbrix compare to Mem0 pricing?

Hebbrix includes features in the free tier that Mem0 charges for, like knowledge graph extraction and multi layer search. Our Pro plan includes smart ingestion and webhook notifications. Check our comparison page for a full feature breakdown.

Ready to give your agents memory?

The free tier includes everything you need to build a real proof of concept. No credit card required.