Mneme

The description states that Mneme is a system designed to give AI agents "a memory they actually own." It includes:

• A sovereign vault for storing memories
• On-chain attestations via Monad blockchain
• A mechanism to keep memory consistent when switching models (e.g., from Claude to GPT)
• GDPR-compliant deletion using a smart contract that issues cryptographic tombstones
• An MCP server for integration with AI tools like Claude Desktop, Cursor, and Windsurf
• A dashboard built with Next.js

The system uses:

• Fastify REST API connected to PostgreSQL with pgvector for semantic search
• Neo4j for temporal knowledge graph
• Redis for write resilience
• Solidity contracts on Monad Testnet (VaultRegistry, AttestationAggregator, DeletionProver, MemoryMarket)
• Integration with Codex for scaffolding and development

Inference The product is a proof-of-concept or early-stage prototype built in one week. It is not evidenced to have been deployed in production or used by end users.

The description states that Mneme was inspired by the problem of AI agents losing memory when switching models, and that current solutions only solve "remember this across sessions" but not "remember this across providers with cryptographic proof you own it."

It positions itself as:

• Sovereign and portable memory infrastructure
• On-chain attestations via Monad
• GDPR-compliant by design
• Enables switching models while retaining context

Inference The positioning is focused on data sovereignty, interoperability, and compliance. It claims to be a foundational layer for AI agents that can be reused across different platforms.

The description does not state who the target customer or ideal customer profile (ICP) is. It implies use cases in enterprise settings due to GDPR compliance features, but no specific customer segments are named.

Not evidenced No evidence of target customer personas, buyer profiles, or use cases beyond a hackathon prototype.

The description does not provide any information on pricing, monetization, or business model. It mentions a "Memory Market" that is live on Monad Testnet but does not state whether it generates revenue or how users pay for memory packs.

Not evidenced No evidence of pricing, monetization strategy, or revenue streams.

The description states:

• Built with Fastify, Next.js, Solidity, Neo4j, PostgreSQL, Redis, Docker, Codex
• Uses viem and Drizzle ORM for smart contract interaction
• Includes an MCP server that works with Claude Desktop, Cursor, Windsurf
• Has a dashboard with demo mode
• Implements batching of attestations to reduce transaction load
• Uses SpaCy NER for PII scanning

Inference The technical stack is modern and includes blockchain components. It appears to be a full-stack prototype built quickly in one week.

The description states:

• Built in one week during OpenAI Build Week
• Live on Monad Testnet
• First real memory pack listing and mainnet launch are planned
• Native plugins for LangGraph and CrewAI are in development

Not evidenced No evidence of user adoption, revenue, or customer traction beyond the prototype.

The description does not mention any competitors. It states that current solutions only solve "remember this across sessions" but not "remember this across providers with cryptographic proof you own it."

Inference The project positions itself as solving a gap in the market for interoperable, sovereign AI memory systems — though no competitive analysis is provided.

• Prototype only: The system is described as a hackathon prototype, not a product with traction or customers.
• No revenue or monetization strategy: No evidence of how the project will generate income.
• Unproven market demand: The description does not show any user feedback, pilot programs, or real-world adoption.
• Limited team size: Only two members (Hassan Rehman, Mrunmayee Daware) are mentioned; no indication of scaling or support structure.
• Testnet only: The Memory Market is live on Monad Testnet, not mainnet — suggesting it’s not yet production-ready.

1. What specific use cases have you identified for enterprise adoption?
2. How do you plan to monetize the Memory Market or other components?
3. Have you conducted any user testing or feedback sessions with AI developers or enterprises?
4. What is your roadmap for moving from testnet to mainnet and scaling the system?
5. Are there any existing partnerships or integrations with AI platforms beyond Claude Desktop, Cursor, and Windsurf?
6. How do you plan to handle data privacy and compliance at scale?