This proposal seeks DAO approval to upgrade the RAG Canister to introduce support for multi-model agent creation and to improve file upload performance through batched and concurrent upload mechanisms.

**Details of the Upgrade **Multi-Model Agent Creation Added functionality for agents to be created with multiple AI models. Enables hybrid inference strategies and broader use cases. Provides flexibility for creators to design more powerful AI agents. Enhanced File Upload Implemented batched file uploads to group files into manageable sets. Introduced concurrent upload support, significantly reducing upload times for larger files or multiple datasets. Improves user experience and scalability for training workflows.

Benefits of This Upgrade Improved Agent Capabilities: Multi-model support allows creators to design more sophisticated and versatile agents. Performance Gains: Batched and concurrent uploads speed up data preparation, making training more efficient. Scalability: Supports larger datasets and more complex workflows as the ecosystem grows. User Experience: Faster uploads reduce friction for creators onboarding new agents. Technical Notes Upgrade Mode: Upgrade (no reset). Migration Impact: None expected — existing agents remain compatible. Security Impact: Low — upload batching and concurrency logic handles only data ingestion paths. Testing: New upload methods benchmarked to confirm faster performance and stable concurrent execution. Requested Action The DAO is requested to approve this upgrade to the RAG Canister, enabling: Multi-model agent creation, and Optimized file upload with batched and concurrent mechanisms. This upgrade directly enhances both the power of AI agents and the efficiency of onboarding data, strengthening ELNA AI’s infrastructure for decentralized AI development. ✅ If approved, the RAG Canister will be upgraded with support for multi-model agents and faster, scalable file upload mechanisms.

This proposal was created using https://ic-toolkit.app

Proposal to Upgrade an SNS Controlled Canister

Target canister: bpsjh-6yaaa-aaaah-adyjq-cai

Wasm info

Embedded module with 1688059 bytes and SHA256 2bfac745ff5a01b9b40e8f17cad61c7d2e89226d8aaebbee153137043075c461.

Mode: Upgrade

Argument info

No upgrade argument.