For the complete organized capability specification, see Unified Feature and Capability Catalog. This catalog distinguishes merged source from planned functionality.
Product vision: A unified private AI environment with Ollama inference, Qwen/Gemma models, Smart Model orchestration, agents, knowledge intelligence, multimodal processing, automation and integrated GoreeCloud services. This is a roadmap, NOT an implemented feature list.
- Resumable Ollama streaming, explicit cancellation, partial-output recovery and concurrency limits
- Multi-conversation safety and concurrency guards
- Model capability inventory, deterministic profile configuration and benchmarking
- Per-user authentication through GoreeCloud Identity
- Scoped workspaces and explicit authorization
- Privacy policies, retention and data export through GoreeShield
- Wardveil defense controls, rate limits and audit events
- Backup, restore, migration, verification and release hardening
- Central GoreeCloud UI components when design tokens are verified
- Real-model validation and evaluation of the new experimental local embedding/vector retrieval API; document parsing, hybrid RAG, OCR, citations and multi-user permissions remain planned
- Hybrid keyword/vector search and evaluated real-document retrieval; synthetic offline metric calculations are implemented but do not establish live-model quality
- Workspaces, project instructions, memory and notes
- File import/export and conversation search
- Voice transcription, synthesis and vision through supported models
- Accessible native desktop/mobile clients by target platform
- Auditable MCP client and per-tool permission system
- Sandboxed tool execution and human approval
- Custom agents, skill packages, visual flow editor and automation jobs
- Subagents, execution budgets, retries, cancellation and tracing
- GoreeCloud Mesh service contracts and authorized integrations
- Web research through allowlisted tools and evidence-based citation
- Collaboration channels and calendars
- Secure embeddable assistants
- Agent developer workspaces and rich interactive artifacts
- Meeting recording/transcription and advanced media processing
- Evaluation dashboards, telemetry, model performance analytics
- Pluggable provider adapters and scaling options
Completion requires code, tests, authorization/privacy review, documentation and runtime evidence. Do not silently mark requirements as complete.