Itsuki
Itsuki is a structured memory layer for AI assistants, coding agents, applications and workflows. It extracts durable facts, events, entities and relationships from conversations and serves that context across connected tools instead of replaying entire chat transcripts.
At a Glance
- Developers building AI applications
- AI assistant and chatbot builders
- Coding-agent users and teams
- Agent-framework and workflow-automation developers
- +1 more
AI Tools by Itsuki
(1)
Itsuki
AI Memory Layer for Agents
Discussions
No discussions yet
Be the first to start a discussion about Itsuki
Latest News
Itsuki Terms of Service became effective; the service is identified as operated by Ziyad Ej in India.
Universal Memory Engine repository added terms-acceptance recording at every sign-in door.
Repository backlog records production go and publication of all packages.
Repository records graph UI polish and preparation for the open-source release.
Products & Services
Hosted memory infrastructure that captures conversations, documents and events; extracts structured nodes, details, dated events and stated relationships; and provides recall, updates, receipts, graph views, exports and deletion.
The open-source Apache-2.0 repository powering Itsuki. It exposes the service through MCP, REST, server-side SDKs, a dashboard and a Claude Code plugin, and is built on Cloudflare.
Python client package for synchronous and asynchronous save, recall/search, status, waiting for enrichment and deletion operations.
First-party JavaScript/Node client for integrating Itsuki memory into applications and agents.
Market Position
Itsuki positions itself against transcript-replay systems and vector-only memory databases by storing typed, source-backed and editable graph objects, combining lexical, semantic and graph retrieval, and returning bounded prompt-ready context. Comparable alternatives in the broader AI-memory space include Mem0 and other MCP memory servers, but Itsuki differentiates around receipts, explicit admission rules, revision history, rollback, exact deletion and a single cross-tool memory layer.
Leadership
Founders
Ziyad Ej
Listed by Itsuki's Terms of Service as the operator; the public materials identify him as the person operating Itsuki from India. No prior roles or companies were identified in the reviewed sources.
Executive Team
Ziyad Ej
Operator
Operates Itsuki from India according to the Terms of Service.
Founding Story
Itsuki was created to address the loss of context between AI sessions and the limitations of transcript replay. Its initial vision is one private, structured, inspectable memory graph shared across AI tools, with receipts, source links, editing, version history and deletion rather than an opaque embedding store.
Business Model
Revenue Model
Hosted SaaS/API model; the service is free during early access, with paid plans potentially introduced later. The product supports usage-based allowances for AI-processed writes and Huba messages, while additional limits can be manually granted by the owner.
Pricing Tiers
The homepage describes the service as free during early access; the docs describe a default allowance of 15,000 Workers AI neurons per day (about 100 direct saves), 1,000 AI-processed writes per calendar month and Huba message allowances.
Target Markets
- Developers building AI applications
- AI assistant and chatbot builders
- Coding-agent users and teams
- Agent-framework and workflow-automation developers
- Organizations needing inspectable, controllable and portable AI memory
- Personalized product experiences
- Conversational continuity across sessions and channels
- Reliable context for coding and planning agents
- Operational memory for queues, jobs, n8n workflows and backend services
- Customer support and adaptive assistants
- Application and agent developers integrating memory through REST, MCP or SDKs