Product update
Connect your AI assistant to Aeglio with MCP
Find overdue tasks, check invoices and turn meeting notes into follow-up work from your AI conversation. Choose the team and actions your assistant can access.
You have just finished a client meeting. Your notes are in an AI conversation, but the follow-up work belongs in Aeglio: a task to review the proposal, a question about an invoice and a deadline to update.
Aeglio’s MCP connection lets a compatible AI assistant work with your Aeglio records from that conversation. You choose the team and the permissions it receives, so you can start with questions and add access to create or update records when you need it.
MCP stands for Model Context Protocol. It gives assistants a way to use tools provided by another application. For Aeglio, that means retrieving information and carrying out supported actions on your behalf.
Start with a question about your work
Instead of copying task lists into a chat, ask your connected assistant to look up the relevant work in Aeglio.
For example:
Use Aeglio to show me overdue tasks in the Website project. Group them by assignee and include their due dates.
Or narrow the question to a particular client:
Find the invoices for North Studio from October. Show their numbers, due dates and outstanding amounts.
These are examples of requests you can make, rather than fixed commands to memorise. Naming the project, client and period helps the assistant find the right records. The information it can retrieve depends on the connection’s permissions and your own access in Aeglio.
Turn a conversation into follow-up work
With the relevant write permissions, the assistant can also create and update records. Task work is a useful place to start: you can move from discussing what needs doing to recording it where your team follows progress.
For example:
Turn these meeting notes into proposed tasks for the Website project. Show me the titles, descriptions and suggested due dates before creating anything.
Once you have reviewed the proposal, you can ask the assistant to create the tasks. If they need assignees, it can first look up the people available for that project.
A smaller request might be:
Find the task “Review the homepage” in the Website project and add a comment: “Client feedback is ready for review.”
Be explicit about whether you want a suggestion or a saved change. After an update, open the relevant records in Aeglio to check the result.
Choose what the connection can do
When you connect, you sign in with your existing Aeglio account and choose a team. You then select the permissions to grant.
Access is split by resource and action. Reading tasks, creating tasks and deleting tasks are separate permissions. Other areas have their own supported actions, such as reading invoices or sending them.
Read-only access is a useful starting point if you want to ask questions without allowing changes. For a task workflow, you can grant the needed task actions while leaving unrelated access unselected.
Your existing team and record permissions still apply. Giving a connection an action permission does not give it access to records you cannot use yourself.
See the access you have granted
Open Profile → Sessions & apps → Connected apps to review a connection and its team. Expand its permissions to see them grouped into blocks such as Tasks and Invoices.
Green Allowed labels show the actions granted to the connection. Red Not allowed labels show supported actions that were not granted. This makes it easier to check, for example, that an assistant can read invoices but cannot send or delete them.
You can disconnect the app from the same page. Disconnection stops future access; it does not undo changes already made. To choose a different set of permissions, disconnect and connect again.
Connect your assistant
Your assistant needs to support remote MCP connections and browser sign-in. Add Aeglio using the connection URL in the guide, then sign in and review the requested access. You do not need to create an Aeglio API key.
The AI assistant connection guide includes setup instructions for supported clients and explains how to choose access and troubleshoot a connection.
Start with one concrete request, such as finding overdue tasks in a named project. Once you are comfortable with the results, choose the additional actions that fit your workflow.