TNB
8f4ed8a306
Review tallyflow-assets-iot (TallyFlow Scout v0.5.3): Collect, Verify, Locate with Edge remote reader / offline CSV locate. Clarify Scout does not replace Core Assets. Refresh Edge, Print parent-container and multi- unit identity notes, Capture share/offline/barcode notes, architecture, use cases, capabilities, and website drafts after src/ repo pulls.
1.8 KiB
1.8 KiB
TallyFlow Edge Review (repo: tallyflow-iot)
Product name: TallyFlow Edge (repo/package still tallyflow-iot)
Tech stack: Rust (Cargo workspace) + Leptos CSR web UI
Branches reviewed: develop through 81fa9d8 (July 2026 pull)
Structure
core/— service/library- RFID drivers: Impinj, Zebra, ATID ATF862 / Invengo XC-RF861 (Master mode)
- Barcode: SNAPI / Aurora / wedge paths
- Remote reader REST API (mobile/handheld submit — used by TallyFlow Scout)
- REST API v1, live tag streaming, reader start/stop/health
- Reader-side filtering (for quotes such as RMS tunnel/filter needs — validate in field)
- Keyboard wedge + optional lookup filter (e.g. Assa Abloy patterns)
- Egress processors: CSV logger, REST, AMQP → TallyFlow
- MCP Streamable HTTP at
/mcp(health, readers, tags; statistics-oriented skills expanded) - Windows Service install path
web/— Leptos UI: reader health, live stream/filter, start/stop, improved statistics display
Key capabilities (market carefully)
- Multi-vendor fixed RFID + remote/handheld readers under one site edge
- Unified tag feed → CSV / REST / AMQP
- Ops UI without vendor tools
- MCP read-oriented copilot surface (not autonomous dock control by default)
- Companion handheld: TallyFlow Scout (Collect / Verify / Locate)
Relevance
- Primary site edge ingestion for docks, tunnels, and multi-reader plants
- Can feed Core via AMQP; can stand alone with CSV/REST and customer apps (e.g. Alice+Olivia console)
- Scout is the supported mobile remote-reader experience for Edge
Next review steps
- Confirm shipped MCP tool list vs docs after statistics/MCP skill commits
- Document ATF862 Master mode hardware prerequisites
- Field-test reader-side filtering before strong marketing claims