A B2B operations platform for businesses that deliver work through partners and across many client sites. It combines an account CRM, a request workflow, project management with real scheduling, and a role-scoped knowledge base — behind one permission model.
This is a working demo, not a mockup. It is populated with a retail rollout scenario — chains, stores and in-store screens — as a worked example; the same structure fits any business that manages clients, sites and delivery teams.
Each role sees a genuinely different system. Sign in as two of them and compare the very same screen.
Every demo account uses the password demo1234 — you can also sign in manually.
Four capabilities that usually live in four different tools, on one account model and one permission model.
A multi-level account hierarchy — organisations, the partners who resell for them, their sites and the assets inside each one. Everything else in the system hangs off this tree, and permissions are expressed in terms of it.
Approval gates, parallel workstreams, and tasks linked by the four standard dependency types with working-day lag. It computes a critical path and locks a baseline, so replanning moves forecasts without quietly rewriting the commitment.
Structured step-by-step guides with prerequisites, screenshots and troubleshooting, authored in the app. Visibility is set per category and per article, and a read-only API exposes the same content to external tooling.
A typed intake workflow with a validated status pipeline, comments, attachments and watchers — on top of eight roles, three enforcement layers, and row-level scoping that narrows what you see without changing which screens you get.
An approved request becomes a scheduled project in one step, carrying its account, contacts and context across.
Which platform account exists, for which client, in which environment — and a record written every time one is revealed.
Delivery health across every territory, drilling down from country to partner to individual site.
The source is private, so here are the decisions worth knowing about — each one is visible in the demo above.
Middleware gates the route, every API handler re-checks the session server-side, and business rules live in named predicates rather than inline role checks. Read and write permissions are deliberately separate functions — so widening who can view the project portfolio can never accidentally grant everyone the ability to rewrite a plan.
Roles decide which screens you reach; scope decides which rows you see. An account executive is bound to a territory and a partner to their own organisation, expressed once as a query filter and applied at the data layer instead of being re-derived in each endpoint.
Tasks link with finish-to-start, start-to-start, finish-to-finish and start-to-finish dependencies, with lag counted in working days so the planner skips weekends. Locking a baseline freezes the planned dates — after that, recalculation can only move forecasts.
A production deployment refuses to boot if its database URL does not match the environment it claims to be. That guard exists because the opposite once happened silently, and a wrong-database incident is far cheaper as a failed deploy than as a support ticket.