Skip to content
Merka.Cloud

Roadmap

Overall goal of Merka.Cloud is to extensible compute/storage/AI overlay over public/private clouds, supporting all business (and family/home) operations. The goal is of course rather ambitious, but everything has to start somewhere. The roadmap also explores the thin separation line between “core” and “extension” functionality, and feedback on that is always welcome. As an example - Analytics often considered an non-functional requirement of a system, leading to potential treatment as Add-On/Extension, however we believe it must be at a core of every component - leading to observability, and ongoing self-optimization/improvements.

Ab Initio - First Setup was always envisioned to be via a Personal Mobile Device (iOS or Android). This gives you externalized 2FA for first user and more critically - external backup of access keys that are harder to reach (semi-online). This may be delayed beyond phase 0.1 simply because of complexity of implementation requirements.

Phase 0.1 - Prototype

  • Core Stack (AutoGPT, LakeFS with AWS S3)
  • Local LLM

Phase 0.2 - Initial Use Case

  • Networking, keys, netdata statistics
  • Versioned Data Model (With Migration/Overlay)
  • Managing API keys for other services, such as AI API, S3 Storage, Compute Clouds, More
  • NoCoDB

Phase 0.3 - Additional Use Cases

  • Personal Information Manager with AI (Hoarder?)
  • Lobe-Chat?
  • Shared Knowledge Base with AI
  • Permissioning in case multi-user multi-device scenario, to allow more granular resource utilization?
  • Cross cloud composition features - running containers remotely (figure out which device it can run on best to suggest)
  • Mozilla SOPS for keys management (and Ab Initio) (also check https://github.com/wemake-services/dump-env and others)
  • Telemetry: (TODO: Decide between OpenTelemetry, NetData, Jaeger, SigNoz, Vector.dev, OpenObserve)

Phase 0.4 - Even More

  • Java conversion to Go, or TypeScript (using OpenRewrite)
  • Conversion between frameworks
  • Automated ToolJet like behaviors
  • Non-Local LLM Models support (with RAG) via API, such as OpenAI
  • More

Phase 0.5 - Merka to Merka connection concepts

Phase 0.6 - Languages and optimizations

  • Support for Co-Routines in relevant Languages
  • Kotlin
  • Java
  • Python
  • TypeScript

This roadmap will mostly exclude extensions, as these are expected to grow organically