The future of workforce management is conversational.
We're building an MCP server for MyGig: connect your AI assistant, and staffing becomes something you say rather than a system you operate. Businesses will request workers and create shifts directly from Claude or ChatGPT. Workers will book shifts directly from their AI agent.
MCP (Model Context Protocol) is the open standard AI assistants use to work with outside platforms. MyGig already runs the hard parts as one system: vetted workers, an award engine that prices every hour with the clause cited, and employment carried end to end. The MCP server exposes that to any assistant that speaks the standard.
- Stage
- In development
- Protocol
- MCP, the open standard
- Auth
- OAuth sign-in, planned
One server, everywhere the standard reaches
ClaudeUse MyGig from Claude on web, desktop and mobile. At launch we intend to publish MyGig in Claude’s connectors directory, so adding it is one click; the same MCP server underneath.
ChatGPTAdd MyGig to ChatGPT as a connector. ChatGPT speaks MCP too, so the same server covers it; we intend to publish the listing when the server ships.+ any assistant that speaks MCP
Work in progress. There is nothing to connect yet.
The MyGig MCP server is in development and no launch date has been announced. Everything on this page describes what we are building, in the order we are building it. Register interest and hear first →
How a booking will run, end to end.
The whole flow in plain requests rather than product screens. The agent prepares every step; anything that books a worker or spends money asks you to confirm before it happens.
- Describe the shift"Two pick packers at the Eastern Creek site, Thursday 6am to 2pm." Location, position and headcount become a drafted shift, or several at once.
- Get the estimate in real timeMyGig runs pre-payroll with full award interpretation on your exact request, so the number that comes back is the real cost: award wages with the clause applied, on-costs, and the flat 15% fee. A quote, not a placeholder.
- Select workers, three waysAuto-select and let MyGig fill it; pick manually from recommended, pre-vetted workers who can work that shift; or invite workers you already know by email. Invited workers onboard themselves and are bookable once vetting clears.
- Review the summary, pick payment, submitOne summary of shifts, workers and cost. Choose your payment method and terms if your account has them, then confirm. This is the human moment: the agent never submits for you.
- Approve timesheets after the shiftWhen the shift is worked, approve or query timesheets from the same conversation. Workers are paid the day after, as always.
The rest of the workforce, same conversation.
- Ask for insights across your shifts and workers: fill rates, spend, hours, in plain language
- Manage shifts and time and attendance
- Manage locations and positions (account setup happens on the web first, for compliance)
- Manage workers and pools: favourites, blocked lists and the people you rebook
- Rate workers or report an issue from the same thread
What stays on the web, at least at first
Business account setup and onboarding stay on the web: they involve sensitive documents, team member details and payment methods, and whether any of that ever moves into an assistant conversation is a compliance decision we will make carefully, not a default. Onboarding through MCP may come later; it is not committed.
Your agent finds the shift. You stay in charge.
The worker side of the same server. Employment, award pay with casual loading, superannuation and next-day pay stay exactly as they are; only the way you find and accept work changes.
- "Find me bar shifts near Newtown this weekend" matched against your tickets and availability
- Accept a shift from the conversation, with the pay shown before you commit
- Ask when your pay lands (the day after every shift, as always)
- Check what an award classification pays before you take the work
An AI agent can only book what a system makes safe.
The compliance sits with MyGig, not the conversation.
MyGig is the Employer of Record: employment contracts, award-interpreted payroll with the clause cited, superannuation and workers compensation are carried by us. An assistant requesting a shift cannot get the employment part wrong, because it was never holding it.
Pricing an agent can actually compute.
One flat, published 15% service fee on wages and on-costs, the same for every client. No negotiated rates and no volume tiers means an assistant can answer "what will this cost" with a real number instead of "contact sales".
Award answers come from an engine, not a guess.
Modern Award rates on this site are generated from the same engine that prices real shifts, and award terms are quoted from the Fair Work award library with the clause cited. That is the data an assistant needs to answer pay questions without inventing them.
Humans confirm the moments that matter.
We are designing the server so an agent can prepare everything, and a person confirms anything that books a worker, spends money or changes someone's employment. Delegation should not mean surprise.
Related: Amy, the workforce AI already works inside MyGig today; the MCP server is how outside assistants will reach the same platform.
Good questions, straight answers.
What is MCP?
MCP (Model Context Protocol) is an open standard that lets AI assistants like Claude and ChatGPT securely connect to outside tools and data sources. A platform that runs an MCP server can be used from inside a conversation: you ask in plain language, and the assistant calls the platform to do the work.
What will businesses be able to do through the MyGig MCP?
Describe a shift (location, position, headcount) and get a real-time estimate: MyGig runs pre-payroll with award interpretation on the request, so the quote is the actual cost. Then select workers three ways (auto-select, manual pick from recommended pre-vetted workers, or invite people you already know by email), review the summary with your payment method, confirm, and approve timesheets afterwards. Insights, shift and worker management follow. Anything that books workers or spends money asks a human to confirm, and the server is in development, so this describes what we're building, not what is live today.
What will workers be able to do through it?
Ask your AI agent to find open shifts that match your tickets, location and availability, accept a shift, and check when your pay lands. Workers are employed by MyGig on award rates with casual loading and paid the day after every shift, and that does not change when the request comes from an agent. This is in development and not yet available.
When does the MyGig MCP launch, and which assistants will it work with?
It is in development and we have not announced a launch date. It is being built on the open MCP standard, so it will work with any MCP-compatible assistant, including Claude and ChatGPT, and at launch we intend to publish MyGig in their connector directories so adding it is one click. Register your interest and we will tell you the moment there is something to connect.
How will authentication work?
The design is OAuth: you sign in with your MyGig account through the assistant’s browser flow, access is scoped to that account and revocable at any time, and no API keys are ever pasted into a chat. Like everything on this page, that describes what we are building rather than what is live today.
What will it cost?
MyGig charges one flat 15% service fee on wages and on-costs, the same for every client, and that does not change based on how a shift is created. No separate MCP pricing has been announced.
Talk to your workforce soon.
Register your interest and you will hear the moment there is something to connect. One update that matters, no spam.