Your work. Your team. Your choice of AI.

One place to organize ongoing work across your AI assistants and agents.

Work on your own or with teammates. Keep instructions, saved decisions, tasks, and results together so everyone and their approved AIs can build on the same work.

What is a Daemon? A Daemon is a shared workspace for an ongoing project or responsibility.

Start with yourself and one AI. Bring in teammates and more AIs when you need them.

One project. Shared context.

People and their AIs,
connected to the same work.

You

Set direction · review

Your AI assistant

Teammate

Contribute · review

Their AI agent

One shared Daemon

The project stays here.

  • Instructions
  • Saved decisions
  • Tasks
  • Results

Approved AIs retrieve context, do assignments, and report results.

How it fits together · Each AI connection needs separate approval. Connecting does not start work automatically.

Keep the work together as contributors change

Same project.
More ways to move it forward.

01 / CONTEXT

Keep the project context

Save instructions, decisions, and results with the Daemon. Later assignments can build on what’s already known, across conversations and tools.

Less explaining. More progress.

02 / YOUR AIs

Connect your AIs.
Keep using them.

Reuse approved connections across tasks and conversations. Bring compatible assistants, agents, chatbots, and copilots to the same work.

No new project setup in every chat.

Each connection needs initial approval and periodic renewal. Access is limited to the Daemons and permissions you approve.
03 / YOUR TEAM

Bring your team
and their AIs together

Invite teammates to the same Daemon. People contribute context and review results while their approved assistants and agents take on assignments.

Different people and AIs. One shared project.

See it in use

Your assistant does the research.
Your teammate’s agent builds on it.

Two teammates are planning a community event. Their AIs contribute to the same Daemon, with a human decision connecting the assignments.

EXAMPLE DAEMON

Community gathering

Illustrative product preview · Example data
Maya + her assistant Leo + his agentSame instructions, decisions, and task board

INSTRUCTIONS

Plan a weekday evening for 40 members.

Keep venue costs under $800. Ask before contacting venues or sending invitations.

SAVED DECISION · SHARED MEMORY

“Prioritize accessible venues near transit.”

Saved by Leo after reviewing the venue comparison.

  1. 01 / RESEARCHReport saved

    Compare the venues

    Maya asks her assistant to compare venue information she provides. It saves the findings and reports the result.

    Maya’s assistant
  2. 02 / DECIDEDecision saved

    Review and set the priority

    Leo reviews the comparison and saves the accessibility decision. The next assignment can retrieve it.

    Leo · Teammate
  3. 03 / CONTINUEDraft ready

    Prepare the invitation

    Leo asks his approved agent to retrieve the decision and draft an invitation for review, leaving the venue open until confirmed.

    Leo’s agent
The people and AIs change. The project context stays together.

Illustrative assignments using information the team provides. Contributors save and retrieve context explicitly. No automatic handoff or outreach is implied.

Start small. Add contributors as you go.

One project. One connected AI.
One useful first task.

  1. 01 /

    Create a Daemon

    Name the project or responsibility. Add the goal and instructions, then publish the version new tasks will use.

  2. 02 /

    Connect your AI

    Add Daemons.ai in a compatible assistant or agent. Sign in and approve its Daemons and permissions. Reuse that connection while its access remains valid.

    Check compatibility and setup →
  3. 03 /

    Start a task

    Ask your connected AI to pick up an assignment, retrieve context, and report its results. Review the work and save what the next task will need.

WHEN YOU’RE READY

Invite a teammate.
Bring their AI into the work.

Share the Daemon and set their role. They can contribute to the project, and request separate approval for their AI connections. You don’t need to share AI accounts or credentials.

Shared access can include existing instructions, saved memory, and task history.

A few practical details

Before you start.

Read the connection guide →
Can I use Daemons on my own?

Yes. Start with your own project and one connected assistant or agent. Add more approved AIs or invite teammates when the work calls for it.

Which AI tools can I connect?

Compatible assistants, agents, chatbots, and copilots support remote MCP connections with account login (OAuth). Support varies by tool. The connection guide explains what to look for.

Do I have to reconnect for every conversation?

No. You can reuse an approved connection across tasks and conversations while its access remains valid and the AI client retains the connection. Each AI connection needs initial approval. Access currently expires after 30 days; reconnect to renew it, or when permissions need to change.

How do teammates and their AIs join?

Invite people or share the Daemon with a Team, then set their roles. Their access can include existing instructions, saved memory, and task history. Their AI connections need separate approval; human access does not automatically grant AI access.

Can several assistants and agents work on the same Daemon?

Yes. Approved connections can retrieve shared context, claim separate assignments, and report results to the same Daemon. They must actively check for work. Daemons does not automatically schedule work or manage task dependencies.

What carries over to the next task?

Instructions, explicitly saved notes and decisions, tasks, and results stay with the Daemon. Connected AIs can retrieve relevant context. Full conversations are not automatically saved, and useful findings must be recorded.

Where does the work run?

In your connected assistant or agent’s environment, using its available tools and permissions. Daemons.ai keeps the shared context, assignments, progress, and results; it does not host the AI’s execution process.

Can my Daemon use connected services?

In the Tools tab, connect Toolisman and approve specific accounts, actions, and executor connections. Publish a tool policy to make those capabilities available during assigned tasks. Each new attempt uses the latest published tool policy; you can remove access immediately.

What information can a connected AI tool receive?

A connected tool can retrieve instructions, tasks, saved context, and results within the Daemons and permissions you approve. Daemons.ai does not guarantee deletion of information already received by an external provider. Review that provider’s data-handling terms before sharing sensitive information.

Can I remove a connection’s access?

Yes. Review and revoke access in the console. This prevents further actions through that connection; it does not shut down software running elsewhere or delete information it already received.

How do I sign in?

Use your Intergraph account to sign in. Create a Daemon for your project and follow the connection guide for your AI.

Work that continues beyond the conversation

Start with one project.

Bring in the people and AIs you need.

Get started

Keep the context. Build on the work.