Skip to main content

Composable, event-driven cognitive memory protocol for AI and multi-agent systems—enabling structured onboarding, symbolic–vector fusion, remix, and full audit trails for semantic continuity and explainable intelligence.

Project description

MeshMem SDK

MeshMem™ is the protocol-native, event-driven SDK for composable, auditable, and lawful cognitive memory in AI and multi-agent systems—built and governed by Consenix Group Ltd.


🚀 What is MeshMem?

MeshMem turns any agent, LLM app, or distributed system into a protocol-governed, semantically continuous, and auditable cognitive agent. It’s more than an “AI memory” layer—it’s the full, lawful substrate for composable, explainable, and capitalizable mesh cognition.


🏆 Key Capabilities

  • Field-complete Cognitive Memory Blocks (CMBs):
    Structured, protocol-auditable memory with symbolic fields and vectors—not just blobs or chunks.

  • Lawful Memory Evolution:
    Every memory change (creation, remix, clarification, canonization) is a protocol event—fully logged, validator/auditor-signed, and auditable.

  • Protocol-grade Semantic Fusion (SVAF):
    Field-wise, mathematically rigorous, drift-bounded fusion; no more “just average the embeddings.”

  • Semantic Drift Detection & Clarifier Logic:
    Built-in drift detection and clarifier agents maintain mesh-wide semantic continuity.

  • Full Auditability & Regulatory Compliance:
    Every event is protocol-logged: who, what, when, why, and how.

  • Composable, Programmable Mesh Memory:
    Remix, clarify, and canonize memories into cognitive assets—ready for assetization, staking, and multi-agent collaboration.


🧑‍💻 Developer Quickstart

Install from PyPI:

pip install meshmem

From source:

git clone https://github.com/meshmem/meshmem-sdk.git
cd meshmem-sdk
pip install -e .

See examples/ for ready-to-run demos.


⚡ Minimal Example

from meshmem import create, remix, clarify

raw_input = "List the main compliance and technical risks for Q4."
cmb, audit = create(raw_input, api_key="sk-...")

# Fuse with anchor CMBs
fused_cmb, remix_audit = remix(cmb, [cmb2, cmb3])

# Clarify semantic drift
decision, clarifier_audit = clarify(fused_cmb, drift_scores={"block": remix_audit["svaf_details"]["block_drift"]}, config={...})

📖 Documentation


💡 Why MeshMem? (vs. Other “Memory” Tools)

MeshMem is protocol-native, mesh-ready, and audit-grade—here’s how it stands apart: (click to expand)

Feature Comparison Table

Feature MeshMem (SDK) LangMem / MemGPT / etc.
Memory Structure Protocol-defined CMBs, fields, vectors, labels “Chunks”, text blobs, vectors
Semantic Alignment Field-wise, drift-bounded SVAF fusion Heuristic, vector avg/concat
Drift Detection Built-in, protocol-governed, clarifier logic Rare or ad hoc
Audit Trail Validator-signed, event-driven, forensic Minimal or app-level
Governance Mesh protocol law, validator circles, RFCs None or app-specific
Assetization Canonize as programmable assets (MCATs, etc.) No, or only as DB records
Compliance Native, protocol-signed, explainable No protocol, no compliance
Mesh/Inter-agent Designed for mesh/collab cognition across orgs Agent-local or per-app

Bottom Line: MeshMem does for AI memory what blockchain did for transactions: It makes memory lawful, composable, auditable, and programmable—enabling collective intelligence, regulatory compliance, and epistemic assetization.


⚡️ Upgrade Path: Open Source → Enterprise

MeshMem is free for non-commercial research, academic, and protocol evaluation. For enterprise/commercial deployments, production support, or integration:

SYM.BOT Ltd — Exclusive Enterprise Steward for MeshMem

Enterprise Support: SYM.BOT

  • LTS releases, onboarding, and SLAs
  • Enterprise integrations, compliance, and security features
  • Professional support, consulting, and managed deployments
  • Custom features and OEM licensing

📧 support@sym.bot
🌐 https://sym.bot

All protocol law and underlying technology remain governed by Consenix Group Ltd.


💼 IP, Licensing, and Commercial Use

MeshMem SDK and protocol are exclusive IP of Consenix Group Ltd.

  • No commercial use, redistribution, or integration permitted without written license.
  • See LICENSE and IP_NOTICE.md for details.

All protocol IP, research, and governance inquiries:
📧 licensing@consenix.com
🌐 https://consenix.com/license


🤝 Contributing

  • Bug reports and documentation suggestions are welcome.
  • Protocol law changes/features require RFC via Consenix governance.
  • All contributions are subject to IP assignment.

See CONTRIBUTING.md


© 2025 Consenix Group Ltd. All Rights Reserved.
MeshMem™, MeshMemory™, CMB™, SVAF™, CAT7™, and SYMBit™ are trademarks of Consenix Group Ltd.


MeshMem: Turn any agent into a protocol-governed, compliant, and composable node of collective intelligence.


Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

meshmem-0.1.0.tar.gz (20.9 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

meshmem-0.1.0-py3-none-any.whl (25.2 kB view details)

Uploaded Python 3

File details

Details for the file meshmem-0.1.0.tar.gz.

File metadata

  • Download URL: meshmem-0.1.0.tar.gz
  • Upload date:
  • Size: 20.9 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.1.0 CPython/3.13.3

File hashes

Hashes for meshmem-0.1.0.tar.gz
Algorithm Hash digest
SHA256 de813f2e4beee2fe60414ace075fcde91ceac5ee2fef8eada8700e6f872ef9bf
MD5 4bd05759fb59fd12456bc237a72b0687
BLAKE2b-256 dfea43eb2a03d6e9cc376cd8e59c428e9765595f6c61677093e27c1b3d05d42c

See more details on using hashes here.

File details

Details for the file meshmem-0.1.0-py3-none-any.whl.

File metadata

  • Download URL: meshmem-0.1.0-py3-none-any.whl
  • Upload date:
  • Size: 25.2 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.1.0 CPython/3.13.3

File hashes

Hashes for meshmem-0.1.0-py3-none-any.whl
Algorithm Hash digest
SHA256 dfe7bf3d1fc8ef5dcf99eff74e5be040010683765b616ab0ebf1f1d7152a21a0
MD5 5afad3b2f7e013125b0d682e35e5bb2b
BLAKE2b-256 7d06cd13a9cd4caad579855cdbde1f10f5cd768e4073ae402d827cad51c1fd11

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page