The Decentralized Model Context Protocol (DMCP) — powered by the Decontext Protocol — is evolving.

Since launch, DMCP has enabled AI agents to interact with blockchains directly—handling transactions, reading on-chain data, and triggering smart contracts. But these agents were stateless. They couldn’t remember what happened yesterday, let alone tailor their behavior based on your preferences.

That changes now.

We’re adding decentralized AI memory to DMCP.

This is a major milestone—not just for us, but for the future of decentralized, intelligent agents.

With the upcoming upgrade, DMCP agents gain long-term memory. This memory isn't stored in centralized servers. Instead, it’s encrypted, distributed, and verifiable.

• Recall past interactions and actions

  Transactions (token transfers, swaps)

  DAO proposals and votes

  Session history across devices and sessions

• Adapt based on personal context

  Preferences, habits, and behavioral patterns

  Your typical DeFi behavior or governance tendencies

• Make smarter decisions, not just follow orders

  Suggest next steps based on historical data

  Prevent repeated mistakes or missed opportunities

• Store all this privately and securely

  – Memory is encrypted client-side

  – Stored on IPFS

  – Proven durable via Crust Network or similar decentralized storage backends

  – Only accessible with your permission

Before this upgrade, DMCP was like a universal remote control—powerful, but unaware of its user. Every interaction was one-off.

Now, it becomes more like a personal AI assistant that evolves with you.

This transforms what AI agents can do in Web3.

From:

- Stateless, command-only bots

- Manual, repetitive DeFi actions

- Generic DAO voting tools

To:

- Persistent, memory-aware agents

- Proactive tools for optimizing DeFi strategies

- Personalized governance companions

Here’s how memory changes your experience:

- “What did I swap on Uniswap last week?”

- “Who did I send 0.5 ETH to on Tuesday?”

- “Remind me when my gas fees go over $50 this month.”

- Track yield farming or LP positions

- Alert when APYs change or rewards are unclaimed

- Understand which protocols you use most—and optimize

- Log all past votes and rationale

- Suggest votes based on past behavior or community trends

- Notify you when proposals match your interest areas

- Watch for unusual transaction behavior

- Notify you if you’re spending more than usual

- Help set and track budget or gas usage limits

• Model Context Protocol (DMCP): Enables LLMs to interact with blockchains

• Powered by: The Decontext Protocol, ensuring modularity, verifiability, and decentralized coordination

• Memory Layer: Encrypted JSON-based context stored on IPFS

• Proof of Storage: Backed by decentralized networks like Crust

• Zero Trust Architecture: Only you can decrypt your memory

This upgrade is fully compliant with decentralized principles—no centralized server.

By adding memory, we move from:

- Temporary tools → Persistent assistants

- Static scripts → Evolving agents

- Centralized LLM backends → Fully decentralized agent frameworks

This upgrade is not just an improvement—it’s the foundation for the next generation of on-chain AI.

We’re finalizing the integration and optimizing performance.

When this goes live:

- Developers can build memory-enhanced AI apps using the DMCP SDK

- Users can try public agents (like AIContext) that grow smarter with use

- Communities and DAOs can build their own governance AI tools

This is the future of AI x blockchain—and it starts here.