Storacha AI Agents Toolkit – March 27, 2025 – Eth Elder Ease – PLDG 2

[seetadev]

Storacha AI Agents Toolkit – March 27, 2025 – Eth Elder Ease – PLDG 2

Contributors

- @seetadev - @Chaitu-Tatipamula - @aspiringsecurity

Accomplished Tasks

• Enhanced Storacha-backed RAG module for Web3 knowledge base by adding support for large medical invoice datasets.
• Addressed libp2p stream reliability issues for large real-time payloads with improved multiplexing logic.
• Deployed improved token-gated access smart contracts for file upload/download on Filecoin FEVM.
• Built context-aware agents with Joule Studio and CrewAI to enable personalized elderly care recommendations.
• Configured a cross-platform AI agent ecosystem with enhanced P2P message routing via Py-libp2p QUIC implementation.
• Refined secure manual KYC onboarding workflows using LIT Protocol and Storacha.
• Mentored team members on improving FAISS indexing and retrieval configurations.

Update

• Eth Elder Ease Integration:

  • Expanded AI agent functionality to include personalized suggestions for caregivers.
  • Improved large dataset support in the Storacha-backed knowledge base.

• Token-Gated Access:

  • Enhanced Filecoin-based token-gated contracts to support role-specific permissions.
  • Tested and validated the tokenized workflow for secure file storage and retrieval.

• libp2p Integration:

  • Enhanced P2P communication for AI agents with improved QUIC handshake stability and better error recovery.
  • Reduced data transfer latency for large payloads using updated stream configurations.

• Challenges:

  • Handling high-volume concurrent FAISS queries with reduced latency.
  • Further optimizing Noise protocol handshake for better cross-platform consistency.
  • Balancing personalization logic for AI agents while maintaining real-time responses.

Demo Video

Please also visit: https://drive.google.com/drive/u/5/folders/1dbxjNMvvJusAuUTA9TietqsXJnI9QiTG for LIT KYC Module's screencasts and demos and https://drive.google.com/file/d/1iW-xNdogm3yLFdzWm1yk9T1PWPgR5XEe/view (Medical Invoice integration with Storacha).

Questions

• Are there recommended methods for integrating tokenized access directly with AI reasoning workflows?
• What techniques can we use to improve the efficiency of FAISS indexing for rapidly changing datasets?
• How can we further reduce latency in libp2p-based multi-agent communication for real-time decision-making?

Next Tasks

• Optimize concurrent FAISS queries with dynamic caching and indexing strategies.
• Improve the AI agent’s ability to manage and process large, complex datasets via CrewAI integration.
• Refine token-gated file storage workflows for seamless cross-platform compatibility.
• Develop a collaborative multi-agent simulation using MetaGPT for elderly care scenarios.

[seetadev]

@jamiechicago312 :

We are encountering an issue while using the fetchUploadProgress feature. When I include the fetchWithUploadProgress option, I get the following error in the console:

Refused to set unsafe header "content-length"

Despite the error, the file does get uploaded successfully. However, this error is consistently logged in the console, and I’m concerned it might indicate an underlying issue that could affect future use cases.

For additional details, I’ve referred to page 7 of the document: AI Agents' Knowledge Base on Storacha - Documentation.

Please let me know if this is a known issue or if there’s a recommended solution to address this error.

Looking forward to your guidance.

Thank you!