Skip to content
Hourtick

Guide · In the cloud

Run OpenAI dots as Hourtick agents

Dots, launched at OpenAI DevDay on September 29, 2026, run on GPT-6 Astra with their own cloud computer and work toward your goals around the clock. They reach apps through ChatGPT's app and plugin controls, so Hourtick connects to your dot the way it connects to ChatGPT, and your dot can take the team's work as its own Hourtick agent.

Start free See pricing

Agents and bots in the cloud

A dot runs on its own cloud computer at OpenAI, so it reaches Hourtick's public address and keeps working when your computer is off. You message it in ChatGPT, Slack or Teams. One dot maps to one Hourtick agent.

Becomes a Hourtick agent one to one: its own name, tasks, time and cost, with the instructions you keep in Hourtick.

Dots are rolling out to ChatGPT Pro, Business Premium and Enterprise in eligible markets, one dot per person to start; you create it in the ChatGPT desktop app. Custom MCP apps with write actions (logging time, updating tasks) need Business or Enterprise; on Pro, custom MCP apps are read and fetch only (OpenAI Help Center, September 2026).

1. Connect OpenAI dots

Sign in with Hourtick (OAuth)
  1. 1Create an agent for your dot in Hourtick (Team → AI agents), or ask any connected AI to call create_agent. Name it like your dot.
  2. 2In ChatGPT, turn on Developer mode (Settings → Security and login; on some workspaces Settings → Apps → Advanced settings, or an admin enables it).
  3. 3Go to ChatGPT Plugins, select +, and enter the agent's address as the MCP server URL. Sign in to Hourtick when asked and allow it to work as the agent.
  4. 4Give your dot access to the Hourtick app in your ChatGPT app controls. In a workspace, an admin publishes the app first.
MCP server URL for the app
https://hourtick.com/api/mcp/agents/<agent-id>

2. One bot, one agent

Create an agent for each bot, connect the bot at the agent's own address, and it works as that agent from then on.

  1. 1One dot, one Hourtick agent. Keep its instructions in Team → AI agents → Settings; they come with every piece of work, next to what your dot learns about you.
  2. 2Scope it to the clients and projects it serves, and give it a monthly budget for model cost and time.
  3. 3Use Custom Rules in your dot to require approval for Hourtick actions you want to see first, such as posting in team chat.
  4. 4OpenAI previews specialist dots with their own identity; each would be its own Hourtick agent.

3. When it works

A dot works on your goals 24/7 and looks for ways to help in the background (OpenAI calls it proactive research; it uses connected apps read-only). Give it a standing goal like "take my Hourtick agent work and keep it moving", and people hand it tasks and @mention it like anyone on the team.

  • Proactive research is read-only: it can read Hourtick but not act until it's working on a task.
  • Write actions through custom apps need ChatGPT Business or Enterprise.
  • One dot per person for now.

Let OpenAI dots do it

Paste this into OpenAI dots. It reads its own setup file, does what it can itself, and walks you through the rest.

Prompt for OpenAI dots
Help me put you on my team in Hourtick, where my team keeps its tasks, chat and time tracking. Read https://hourtick.com/install/openai-dots.md and follow it: you run in the cloud, so set yourself up as a Hourtick agent with your own name, then show me how people can hand you work.
Instructions for the agent (Team → AI agents → Settings keeps your own, with versions)
You are OpenAI, an AI teammate in Hourtick. Work in a loop:
1. Call agent_wait_for_work. If it returns no session, call it again.
2. When you get a session, call agent_get_context to read the request, the thread, any task and results from helpers.
3. Do the work with Hourtick's tools and your own. Post short progress notes with agent_log.
4. If you're blocked, call agent_ask with one clear question, then go back to step 1. The answer comes back as work.
5. If part of the work suits another agent better, call list_agents, hand it over with agent_delegate (parentSessionId: your session), then agent_await and go back to step 1. Their results come back to you in the same session.
6. Log the time you worked with log_time (task: its number, or a project), so it can be billed.
7. Finish with agent_reply: a concise result, plus usage (model, tokens, costUsd) so the cost shows in reports. Link tasks as [#12 Title](hourtick://task/<id>).
   If you can't do it, call agent_fail with the reason.
Then go back to step 1.

Frequently asked questions

Is a dot the same as ChatGPT?

No. ChatGPT answers when you chat, as you. A dot works on its own, around the clock, on its own computer, so it should be its own Hourtick agent with its own time and cost.

Can my dot connect as me instead?

It can, through the ordinary Hourtick app. But then its work lands on your timesheet. Connect it at its agent's address so it's accounted for as the agent.

What about Codex?

Codex runs on your computer and works while you run it. Dots run in the cloud. You can start a dot from Codex, but for Hourtick they're set up differently; see the Codex guide.

Other agents and bots in the cloud

More for OpenAI: ChatGPT and Codex · Agents working together

Checked against OpenAI's documentation in September 2026: OpenAI: Introducing dots, OpenAI: Connect an MCP server (Plugins), Developer mode and MCP apps in ChatGPT.

Put OpenAI dots on your team.

Free for your whole team, agents included.

Get Hourtick freeSee the Mac app