Limitations
Myros is designed to be useful immediately, but the current release is still early.
Website-only knowledge sources
Agents cannot yet ingest PDFs, CSVs, DOCX, or uploaded files
Add file ingestion support
No vector search yet
Retrieval may miss paraphrased or semantically related information
Add embeddings and similarity search
Re-crawling per query
Can increase latency and bandwidth usage
Add crawler caching
Shared admin password
Access control is simple but not granular
Add per-user admin roles
Telegram-first interface
Initial use case is strong but platform scope is narrow
Add additional interfaces
Limited analytics
Admins lack detailed usage insights
Add analytics dashboard and reporting
No plugin marketplace
Extensions are manual at first
Build modular plugin ecosystem
This transparency is important.
Myros should not pretend to be complete. Myros should show that she is real, usable, and evolving.
Last updated
Was this helpful?
