A mission control for ChatGPT Work is the layer that turns its one-agent-per-task model into a team: many named agents on one standing backlog, waking each other instead of waiting for the next prompt. ChatGPT Work itself is the strongest one-outcome agent OpenAI has shipped, and one-outcome is exactly its boundary.
3-15
active Scheduled Tasks per account is OpenAI's cap, plan-dependent, at a minimum one-hour cadence with auto-pause on inactivity. A business with more lanes than its cap has to choose which lanes exist.
Source: OpenAI, Scheduled Tasks documentationiShort answer
ChatGPT Work (July 9, 2026, on GPT-5.6) takes an outcome, gathers context from your connected apps, and works for hours until it ships a finished document, spreadsheet, deck, or app. It is one agent per task by design: subagents cannot message each other, scheduled tasks are capped per account, and visibility is per task. Continuous business lanes need many agents on one backlog, and that is MissionControlHQ: a persistent squad at $99/mo flat plus the flat-rate AI plan it runs on, while ChatGPT Work stays your deliverable machine.
Key takeaways
| Question | Short answer |
|---|---|
| Is ChatGPT Work multi-agent? | One agent per task; subagents are hierarchy-only with no peer messaging |
| How many schedules can it hold? | 3-15 active per account by plan, one-hour minimum cadence |
| Can agents hand work to each other? | No: a human routes every output to the next input |
| Whose budget does it drain? | The shared ChatGPT credit pool on your account |
| What persists in a squad? | Task board, threads, memory, schedules, cost history, for months |
ChatGPT Work ships one finished thing per task. A mission control runs the lanes that never finish.
ChatGPT Work owns the deliverable
Docs, sheets, decks, and apps, produced by one agent working for hours with plan-mode and check-ins.
One agent, one task, by design
No peer messaging between agents, capped scheduled tasks, per-task visibility only.
Business lanes are relays
Research feeds writing feeds publishing feeds follow-up: outputs must become the next agent's inputs.
The squad layer runs relays
MissionControlHQ: named agents, shared backlog, mention-driven wakes, runs ledger, live share link.
What ChatGPT Work is, and what it is scoped to
ChatGPT Work is OpenAI's work agent, released July 9, 2026 on GPT-5.6: give it an outcome, and it gathers context across connected apps, breaks the job into steps, and works independently for hours until it ships a finished document, spreadsheet, presentation, or interactive app (publishable via Sites). Trigger-based tasks can even turn new Slack or Teams messages into updated docs.
The scope is one agent per task, by OpenAI's own documentation:
- Subagents are hierarchical. A task's agent can spawn hosted subagents, but they communicate strictly parent-to-child with no peer-to-peer messaging.
- Workspace Agents are configs, not colleagues. Enterprise/Business shareable agents are named configurations humans chat with, with no inter-agent messaging or shared state between agents.
- Schedules are capped. Scheduled Tasks run 3-15 active per account depending on plan, minimum one-hour cadence, auto-paused on inactivity.
- Visibility is per task; organization-level analytics are Enterprise admin features. Billing draws from the account's shared credit pool (help article).
For shipping a deliverable, one strong agent is the right design. For operations, it is the exact gap.
The relay problem one agent can't solve
The relay problem is what business lanes actually are: research feeds writing, writing feeds publishing, publishing feeds follow-up, and support feeds all of them. Each handoff in ChatGPT Work has exactly one router available: you. The research task finishes; a human reads it, decides, and describes the writing task. Multiply by every lane, every day.
The caps make the ceiling concrete. A founder running support checks, content, lead follow-up, competitor watching, invoicing, and reporting needs more standing schedules than a per-account cap of 3-15 allows, so lanes get rationed. And because visibility is per task, the question "what did all my agents do this week and what did it cost" has no screen to answer it.
What to look for in the layer above tasks
The layer above tasks earns its place by running relays without a human router. Five criteria, in priority order:
- Persistent named agents. The same specialists, with their memory, exist next month without re-briefing.
- One durable backlog. A task board every agent and the founder read and write, with claiming, threads, and history.
- Agent-to-agent wakes. Mentions, schedules, and inbound email trigger the right agent directly; outputs become inputs without a human ferry.
- Uncapped, squad-owned schedules. Lanes are not rationed against a per-account allowance, and their costs are attributed per run.
- A shareable live view. A co-founder or operator can watch the whole squad without your login.
What a mission control adds to ChatGPT Work
MissionControlHQ is a hosted mission control built against those five criteria:
- Named specialists that persist. The squad is designed in a Telegram conversation with a lead agent; the same agents, with their memory, exist next month.
- One shared task database. Agents claim tasks off a board, discuss them in threads, and archive them with full history, on the same board the founder watches.
- The relay, built in. An @-mention in squad chat triggers the mentioned agent's run; so do cron schedules and inbound email (agent inboxes are a paid add-on). The research agent's document wakes the writing agent directly.
- Founder-grade visibility. An activity feed across every agent, a runs ledger with per-run cost, model, and trigger, and a live read-only share link.
- Its own subscription underneath. The squad runs on a connected flat-rate plan, so your ChatGPT credit pool stays yours.
They compose cleanly: ChatGPT Work for the big one-off deliverables, the squad for the lanes between them. The full landscape comparison covers all seven tools; siblings of this guide cover Claude Cowork, Codex, Claude Code, OpenClaw, and Hermes Agent.
What the squad stack costs
With published prices as of July 2026, pre-calculated:
- ChatGPT Work: included in paid ChatGPT plans (pricing), billed from the account's shared credit pool; heavy agentic use points at the $100-200 tiers.
- MissionControlHQ: $99/mo flat plus the flat-rate AI plan the squad runs on, with the $100-200 tiers recommended (a $20 plan's limits run out almost immediately under squad workloads). Typical all-in: $199-299/mo, no token markup, no per-agent seats.
- The honest comparison is not plan-vs-plan but the alternative way to run continuous lanes: a junior ops/marketing hire at roughly $4,000/mo.
~94% less than one junior hire
A full squad at $199-299/mo all-in ($99 + a $100-200 flat AI plan) vs ~$4,000/mo for a single junior ops hire: roughly $44,000-45,600 saved per year, with your ChatGPT credit pool untouched.
How to choose
What shape is the work?
- If one big deliverable: a deck, a sheet, an app→ChatGPT Work
- If a few personal recurring checks within the cap→ChatGPT Work Scheduled Tasks
- If relays and lanes owned by specialists→MissionControlHQ squad
Who routes outputs to the next step?
- If you, and that's fine at your volume→stay inside ChatGPT Work
- If agents should wake each other→MissionControlHQ's mention-driven relay
Who needs the whole picture?
- If just you, task by task→ChatGPT Work's per-task views
- If you across weeks, or a co-founder→MissionControlHQ's feed, runs ledger, and share link
Use-case cheat sheet
| Scenario | Best pick | Why |
|---|---|---|
| Board deck from a messy quarter of files by tomorrow | ChatGPT Work | One outcome, hours of autonomous work, finished artifact: its exact design. |
| Slack thread → updated spec, automatically | ChatGPT Work triggers | Trigger-based tasks are its strongest always-on feature. |
| Six lanes running weekly without rationing schedules | MissionControlHQ | Squad-owned crons without the 3-15 per-account cap. |
| Research → draft → publish relay without you routing | MissionControlHQ | Mentions wake the next agent; Work's subagents cannot message each other. |
| Know the month's agent spend per lane | MissionControlHQ | Runs ledger attributes cost per run; Work's analytics are Enterprise-gated. |
| Publish a polished microsite from a prompt | ChatGPT Work (Sites) | Sites publishing is native and shareable by URL. |
When ChatGPT Work alone is the right answer
ChatGPT Work alone is the right answer when the work is genuinely outcome-shaped:
- Each job ends in one finished artifact you can describe up front.
- Your recurring needs fit inside the scheduled-task cap with room to spare.
- You are the only consumer of the results, and per-task visibility is enough.
Outcome tools for outcomes, a squad for the lanes. Most founders end up with both, and they compose: ChatGPT Work ships the deliverables while the squad runs the relays around them.
Frequently asked questions
Task mechanics
Is ChatGPT Work multi-agent? Not in the persistent sense. It runs one agent per task, which can spawn hosted subagents that communicate strictly parent-to-child with no peer-to-peer messaging. Enterprise Workspace Agents are shareable configurations humans chat with, with no inter-agent messaging or shared state between agents.
How many scheduled tasks can ChatGPT Work run? OpenAI caps Scheduled Tasks at 3 to 15 active per account depending on plan, with a minimum one-hour cadence and auto-pause on inactivity. A business with more lanes than its cap has to choose which lanes exist.
Does ChatGPT Work show what all my agents did? Visibility is per task: each task has its own activity view. Organization-level analytics are Enterprise admin features. There is no cross-business feed answering what every agent did this week and what it cost.
Fit and pricing
Is MissionControlHQ a ChatGPT Work replacement? No. ChatGPT Work is the strongest one-outcome agent OpenAI ships, and it stays that. MissionControlHQ is the layer above tasks: many named agents working one standing backlog, waking each other with @-mentions, on schedules without per-account caps that force choosing between lanes.
Can I run MissionControlHQ on my ChatGPT subscription? Yes. Most founders connect a flat-rate ChatGPT plan, with the $100-200 tiers recommended for squad workloads (a $20 plan's limits run out almost immediately). Claude with Extra Usage, MiniMax, and Z.AI also work. MissionControlHQ adds no token markup.
What does the squad cost next to ChatGPT Work? ChatGPT Work is included in paid ChatGPT plans and billed from the shared credit pool. A MissionControlHQ squad is $99/mo flat plus the recommended $100-200 flat AI plan, so $199-299/mo all-in: roughly 93-95% less than the ~$4,000/mo junior hire the same continuous lanes would otherwise need.
What does a ChatGPT Work user actually gain from a mission control? A relay instead of a queue of one. In a squad, the research agent's output wakes the writing agent, drafts wait at approval gates, schedules run without per-account caps, and one live dashboard shows every agent's runs and costs, shareable by link.
Sources
- OpenAI: ChatGPT Work announcement, Work and Codex billing, Scheduled Tasks, Workspace Agents, subagents docs, ChatGPT pricing
- MissionControlHQ: homepage, early access
- Related on this site: Why MissionControlHQ, Mission Control for Claude Cowork, Mission Control for Codex
Last updated: July 2026. Pricing and features verified as of July 2026.