Open Tomorrow
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@Open TomorrowCheck Saturday's shop handover plan for coverage conflicts."
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Open Tomorrow
The last check before your day off.
A fictional flower-shop owner has arranged Saturday cover. One helper is assigned to set up the shop 08:00–09:00 and deliver flowers 08:30–09:15. That is not a workable handover. Open Tomorrow exposes the exact conflict, separates assignment from acceptance, and reopens the affected duty when a helper cancels.
This is an early, working local MCP server, not a completed contest entry or Alexa integration. Its helpers and acceptance actions are explicitly simulated. It does not send messages or alter anybody's real schedule.
Run the complete proof
Use Node 22+ and npm:
npm ci --ignore-scripts
npm test
npm run demoNo API key, AWS account, Amazon device, model service, payment, or sign-in is required. demo starts a server on an ephemeral loopback port, connects using the official MCP client over Streamable HTTP, discovers and calls the actual tools, restarts the server, and checks that state survived. Fixture state is saved to a uniquely named OS temporary directory; no user plan is overwritten.
The successful demonstration proves:
The exact 08:30–09:00 overlap blocks coverage.
Moving delivery to a second helper does not invent acceptance.
Simulated helper acceptance plus explicit entered-scope confirmation permits the narrowly scoped result, “All entered tasks have simulated accepted coverage.”
Cancelling the second helper reopens delivery while leaving setup and its acceptance intact.
Acceptance from the cancelled assignment is rejected.
The changed plan survives a process restart and a new MCP connection.
Related MCP server: MCP-Based Personal Productivity Agent
MCP surface
npm start serves http://127.0.0.1:8917/mcp and stores its single local plan in .local/plan.json. Compatible agent clients can call:
create_handover: initialize explicit dated tasks and helper availability once.check_coverage: read exact blockers, entered-scope status and last affected tasks.change_handover: version-checked assignment, simulated acceptance, cancellation, time correction, or scope confirmation.
Protocol 2025-11-25, negotiated and verified in the demonstration using @modelcontextprotocol/sdk 1.30.0. This is not a wrapper claiming integration only in documentation: the actual client/server transport and tools run in the proof.
Deliberate boundaries
All acceptance is a simulated helper action, not identity-authenticated consent. A production acceptance flow is not implemented.
The result covers entered tasks only. Missing duties are unknown, not safe. This is not labor-law, staffing, travel-time, safety or business-continuity advice.
One day, explicit local wall-time minutes, one operator/server process. Multiple processes sharing a store are unsupported. No shift import, optimization, real-time integrations, employee messaging or automatic plan completion is claimed.
Loopback only; browser origins rejected. Do not expose this development server to the network. Authentication, multi-user isolation and public hosting are not implemented.
No language-model reasoning, speech recognition, Alexa-device execution or end-user validation has been demonstrated yet. The scripted proof is labeled as a scripted proof.
Contest status and provenance
Candidate for the Alexa+ track of the Amazon Developer Hackathon. Not registered and not submitted as of September 11, 2026. The prior Open Tomorrow idea was written in July; this implementation began September 11. Final entry must disclose this history accurately and include a judge-facing experience, public under-three-minute video and truthful tool feedback. Provider confirmation, not this README, decides submission status.
AI-assisted design and code by Codex. The authored fixture contains no real shop, employee or customer information. No third-party media assets are used. The MCP SDK and Zod retain their respective open-source licenses.
This server cannot be deployed
Maintenance
Related MCP Connectors
Shared task queue for humans and AI agents: leases, handoffs, approvals and signed receipts.
Staff scheduling — manage staff, shifts, assignments, certifications, and requests via AI.
Generate optimized work shift schedules from staffing demand, availability and fairness rules.
Field workforce scheduling for AI agents. GPS punches, forms, timesheets. No dashboard.
Related MCP Servers
- AlicenseNot gradedqualityBmaintenanceProvides MCP tools for operations desk tasks including calendar availability, customer lookup, quote calculation, and notification sending. Includes an internal agentic orchestrator that consumes the same MCP tools via protocol.MIT
- FlicenseNot gradedqualityBmaintenanceEnables personal productivity management by exposing MCP tools to create, retrieve, list, update, complete, and delete tasks and calendar events, with validation and conflict detection.-
- AlicenseNot gradedqualityBmaintenanceEnables AI coding agents on a team to coordinate by declaring plans, checking them against other in-flight plans and shared memory, receiving proceed/wait verdicts, and recording handoffs and rulings.MIT
- AlicenseNot gradedqualityAmaintenanceEnables MCP agents to help a person understand their goals, discuss obstacles, plan actions, and record progress in their own words, with a scoped API client and human-led workflow.AGPL 3.0