Can Memori use my own database?

Yes. Memori supports bring-your-own-database, and data stays in your environment unless Memori Cloud is explicitly enabled.

Is there a free option?

Yes. There is a free open-source self-hosted version and a Cloud Free plan with 5,000 memories created and 15,000 recalled.

Which LLMs does it work with?

Memori is described as LLM-agnostic.

Is there a program for startups?

Yes. The startup program offers free access to the Pro plan for 3 months.

GibsonAI

Memori is an LLM-agnostic memory layer that gives AI agents structured, persistent memory with targeted, low-token recall.

Category:

Optimization

Maker:

Pricing

Paid

Skill Level

Developer

Use Cases

SaaS

⚡️ Tags detected by AI

8 visitors went to GibsonAI from LandinTools in the last 30 days.

Your product? Claim it to see where they come from

About GibsonAI

Memori, from Memori Labs (the site at the GibsonAI address), is an LLM-agnostic memory layer for AI agents. It captures each chat turn and agent execution and classifies it into structured, persistent state such as facts, preferences, rules and summaries, then recalls only the relevant context when a prompt needs it. Semantic search enrichment, a memory graph, and per-result explanations of why each item was included are part of the product.

Developers add the SDK to existing code with one line, and production plans offer a drop-in proxy. Memory can be stored in the team's own database or in Memori Cloud. Controls include role-based access, SSO and OAuth integration, retention windows per namespace or user, audit trails and analytics on memory creation, recall and cache hit rate. The company reports 87% accuracy on the LoCoMo benchmark with 97% lower token usage than full-context retrieval.

Memori is aimed at developers and enterprises running production AI agents. Testimonials on the site come from people at MongoDB, Cockroach Labs, DigitalOcean, SK Networks, Swarms and Agno.

Best for

Developers building AI agents that need persistent memory
Teams reducing LLM token costs in production
Enterprises needing audited, access-controlled agent memory

Pros

Open-source self-hosted option with bring-your-own-database support
Reported 87% LoCoMo accuracy with 97% lower token usage
Explainable recall showing why each result was included
RBAC, retention policies, audit logging and SSO support

Cons

Production plans start at $60K per year
Free cloud tier capped at 5,000 memories created and 15,000 recalled
Paid plans require contacting sales

Pricing

Free open-source self-hosted and Cloud Free ($0, 5,000 memories created and 15,000 recalled) tiers, with production plans from $60K/year (Team), from $150K/year (Business) and custom Enterprise pricing.

GibsonAI FAQ

Can Memori use my own database?: Yes. Memori supports bring-your-own-database, and data stays in your environment unless Memori Cloud is explicitly enabled.
Is there a free option?: Yes. There is a free open-source self-hosted version and a Cloud Free plan with 5,000 memories created and 15,000 recalled.
Which LLMs does it work with?: Memori is described as LLM-agnostic.
Is there a program for startups?: Yes. The startup program offers free access to the Pro plan for 3 months.