Mengram
Provide persistent, multi-type memory (Semantic, Episodic, Procedural) for AI agents so they never forget and continuously improve.
At a Glance
- AI Developers
- Enterprise DevOps
- Autonomous Agent Builders
AI Tools by Mengram
(1)
Mengram
Cognitive Memory API for AI Agents
Discussions
No discussions yet
Be the first to start a discussion about Mengram
Latest News
Mengram Launches v2.23.0 with Full Self-Hosting Support
Mengram Introduces Unlimited Self-Hosted and Business Pricing Plans
Integration with LangChain and CrewAI released in v2.5.4
Deepgram acquisition of YC AI startup (Note: Contextual research suggests Mengram remains independent while others are acquired)
Products & Services
Persistent memory API for AI agents covering Semantic, Episodic, and Procedural memory types.
A suite of agents (Curator, Connector, Digest) for automated memory cleaning, pattern detection, and weekly briefs.
Self-hostable version of the memory system for on-premise deployment and privacy-focused use cases.
API that returns a ready-to-use system prompt combining all memory types for a given user.
Market Position
Focuses on a three-type memory model (including procedural) compared to vector-only memory stores like Pinecone or Mem0.
Leadership
Founders
James Kelly
Ex-Google Project Manager (Google Fiber, Android, Maps). Founder of Hijimbo and Mengram.ai.
Ali Baizhanov
Software Developer, Technical Founder. Previously at Prime Source LLP. Creator of the Mengram open-source project.
Executive Team
James Kelly
Founder & CEO
Ex-Google Project Manager for Google Fiber, Android, and Google Maps.
Ali Baizhanov
Technical Founder / Lead Developer
Lead architect of the Mengram open-source repository and managed API.
Founding Story
Started to fix the 'forgetfulness' problem of AI agents between sessions by creating a memory system that auto-captures facts, events, and workflows.
Business Model
Revenue Model
SaaS Subscription and API-usage based pricing.
Pricing Tiers
40 adds, 200 searches, 3 agent runs, 3 sub-users.
100 adds, 500 searches, 10 sub-users, webhooks.
1,000 adds, 10,000 searches, LLM reranking, smart triggers.
3,000 adds, 20,000 searches, unlimited runs, Cohere reranking.
8,000 adds, 30,000 searches, unlimited teams and sub-users.
Dedicated infrastructure, SSO, data residency, priority support.
Target Markets
- AI Developers
- Enterprise DevOps
- Autonomous Agent Builders
- Autonomous AI Agents
- Coding Assistants (Claude Code, Cursor)
- DevOps Automation
- Customer Support Personalization
- OpenClaw Users
- CrewAI Developers
- n8n Community