Connectors let Enterpret Agent read context from and take action in the other tools your team uses. Wire up Slack, Linear, Notion, and the rest, and the agent can pull a spec from a doc, cross-reference usage analytics, post a message, or file a ticket without you leaving the conversation.
Connectors authenticate per user. Each person connects their own accounts, and when the agent acts through a connector it acts as you, with your access and your approval.
What connectors do
A connector is a live link between Enterpret Agent and one of your other tools. It works in two directions.
Reading context in. The agent can search and read across a connected tool: meeting transcripts in Granola, product analytics in PostHog, support conversations in Intercom, designs in Canva. This pulls live context from outside the Context Graph into the work the agent is doing.
Taking action out. When the work calls for a change in another system, the agent writes it there. It posts a Slack message, files a Linear issue, comments on a Jira ticket, updates a Salesforce record. The action lands in the real tool, attributed to you, and the agent asks before it makes the change.
The two directions combine into loops the agent runs end to end. Read a feature spec from Notion, mine the Context Graph for customer evidence, write an evidence section back into the Notion doc, then send the spec's author a Slack message with the key quotes. The agent reads from one tool, reasons over your feedback, and acts in another, all in a single session.
The connector catalog
The catalog grows continuously. The apps below are live today.
App | What the agent can do |
Slack | Post messages, search conversations across the workspace, read channel and thread history, add reactions |
Linear | Create issues, search, comment, and manage projects |
Atlassian Jira | Create issues, search projects, manage tickets |
Notion | Search the workspace, read pages, manage databases |
PostHog | Query product analytics, read insights, manage feature flags |
Intercom | Read support conversations, send messages, manage contacts |
Amplitude | Query analytics, read user behavior data |
Calendly | Schedule meetings, query availability, manage event types and booking links |
Canva | Search designs, read assets, manage content |
Zoom | Search meetings and content, retrieve recordings and assets, read Zoom Docs |
Salesforce | Search accounts, opportunities, and contacts, read and update CRM records |
Granola | Search meeting notes, retrieve transcripts, summarize meeting context |
GitHub | Read repositories, manage issues and pull requests |
If an app you use is not listed, reach out to your Enterpret team. New apps ship on a rolling basis.
How connectors authenticate
Connectors use a remote MCP server with managed OAuth. Connecting one opens an authorization window for that app. Sign in with your own account, grant access, and the connection is live. There is no token to copy or configuration file to edit.
Connecting an app
In a session, open the plus menu in the command box and select Connectors.
Find the app you want in the catalog.
Click Connect. An authorization window opens for that app.
Sign in with your own account for that tool and grant access.
The connector returns to the panel marked as connected and is ready to use.
You connect each app once. The connection persists across your sessions until you disconnect it.
Turning connectors on and off per session
Connecting an app makes it available. Whether the agent uses it in a given session is yours to control from the Connectors panel in the command box. Enable the connectors relevant to the task and leave the rest off, so the agent works with the right tools and nothing it does not need.
Open the panel before you type to see what the agent is wired into, then toggle connectors on or off for the work at hand.
Requesting a new connector
The catalog grows from demand. If the tool you need is not live yet, reach out to your Enterpret team and we will factor it into what we build next.
FAQs
Do connectors get set up per user, or can an admin configure them for the whole org? Per user. Because a connector takes on your identity when the agent acts in a third-party tool, each person connects their own accounts. That keeps every action attributable to you, scoped to your real access, and individually approvable.
Can the agent send a message or file a ticket without my approval? In the default permission mode, the agent asks before it writes to any external system and waits for you to approve. In full access mode, it acts autonomously. Either way, you see a live trail of every action, including connector calls.
The tool we use is not in the catalog. What are our options? Reach out to your Enterpret team, and the agent can work over an uploaded file in the meantime. You can upload flat files (CSV, text, data) directly into a session and have the agent analyze them alongside the Context Graph.
Have a question about connectors, or an app you want to see added? Reach out to your Enterpret team.
