Support and engineering see the same work
A conversation becomes a linked Linear or Jira issue, and the two stay in sync. Never duplicated.
Paste the Sentry link into a ticket by handConnect your code, your error tracker, your issue tracker and the places your team talks. Tore reads what each one already knows, finds the cause in your code and prepares the fix for your engineer to approve.
Tore opens soon. Join the waitlist and we'll email you the day you can start.
Priya · replay
Dan · console
Lena · networkEach connection removes a step where a person copies something a machine already had.
A conversation becomes a linked Linear or Jira issue, and the two stay in sync. Never duplicated.
Paste the Sentry link into a ticket by handGet notified, reply and act without a second tab open all day.
Keep the help desk open in another tabConnect GitHub and every bug is traced to the file, the line and the commit.
Walk an engineer through the bug firstImport your conversations from Zendesk or Intercom and keep every customer's past.
Start over with an empty inboxGrouped by the job the connection does. Pick a group to narrow the list.
Built in. Customers chat from your app, and a bug report arrives with the replay, console and network log attached.
Forward your support address in. Replies land on the right thread and attachments are scanned.
Investigations read the real file and commit. The fix arrives as a pull request for your engineer to approve.
The same investigation on your GitLab projects, with the fix prepared as a merge request.
Bitbucket repositories are read the same way, so the cause is found in your actual code.
Deploys line up against new errors, so a regression points straight at the release that caused it.
Azure Repos are read the same way, so the cause is found in the code you actually ship.
Failed builds and test runs sit next to the errors they line up with.
Errors group by fingerprint and can open an investigation on their own, before anyone writes in.
Use the error tracker you already run. Datadog errors feed the same grouping and investigations.
Metrics and logs pulled read-only as evidence, without giving anyone shell access.
Rollbar errors feed the same grouping and investigations as everything else.
Keep the crash reporting you already run. Bugsnag errors open investigations too.
A conversation becomes a linked issue without retyping, and the two stay in sync.
Bugs land in Jira with the evidence attached. Closing the issue updates the customer.
Turn a conversation into a ClickUp task with the evidence attached, no retyping.
Bugs become Asana tasks linked to the conversation they came from.
Get notified, reply to customers and act on the queue without leaving Slack.
The queue and its alerts reach your team in the channels they already use.
Alerts and handoffs post to your server, so nobody has to watch a second tab.
Customer context from HubSpot arrives with the conversation, so nobody goes looking for it.
Account details sit next to every conversation, pulled from the records you already keep.
Import your Zendesk history, so switching does not mean starting from an empty inbox.
Bring your Intercom conversations across and keep every customer's history.
Account details from Attio arrive with the conversation, so nobody goes looking.
Bring your Front conversations across and keep every customer's history.
Tore's MCP server puts your inbox and investigations inside Claude. Ask, and it answers from Tore.
Pull a bug's evidence and investigation into Cursor through the MCP server while you code.
Any AI tool that speaks MCP can read your inbox and investigations.
Something missing? Use the API and webhooks to connect anything else.
Across your stack
The error arrives from Sentry. Tore links it to an issue in Linear, tells the team in Slack and prepares the fix on GitHub. Each tool shows the same story, in the place people already look.
ENG-412Report export stops after the first page
AlexENG-412: cause found at paginate.ts:57. 3 customers affected. A fix is ready for review.

Looking now.
Approved by AlexSlack
New conversations, handoffs and fixes waiting for review post to the channel you choose. Reply to the customer or take the conversation from the message itself.
Maya at Northwind can't invite teammates. I'm not confident on this one, so I've drafted a reply for a person to check.

The invite button is greyed out for me.
Invites need the Admin role. Your workspace owner, Tom, can switch it on under Settings, Members.

Sent. Reply delivered to Maya in chat
Linear and Jira
A bug from support becomes a linked issue with the replay, console and network log attached. When the issue moves, the conversation knows. When it closes, the people who reported it are told.



You own every connection. You can see what Tore can reach and take it back whenever you like.
You stay in control
GitHub, Sentry, Linear, Slack and email connect from one settings page. The GitHub app reads your repository and opens proposed fixes as pull requests on its own branches. It never merges or pushes to your main branch, and one click takes that access away again.
Webhooks, a scoped API and an MCP server, so Tore reaches whatever you already run.
API, webhooks and MCP
Webhooks push events out the moment they happen. The API reads and writes conversations, issues and more. The MCP server puts your inbox and investigations inside Claude, Cursor and other AI tools.
// POST https://hooks.northwind.dev/tore x-tore-signature: sha256=9f2c…e41a { "type": "investigation.fix_ready", "issue": "ISS-311", "reporters": 3 }
# test-mode key, scope: conversations:read curl https://api.tore.ai/v1/conversations?status=open \ -H "Authorization: Bearer tore_test_••••7Qx"
{ "data": [
{ "id": "conv_8812", "customer": "Northwind",
"status": "open", "assignee": "tore-ai" },
…
] }ISS-311 export stops after the first page. Fix waiting for Alex's approval.ISS-307 invite button disabled for members. Draft replies sent.ISS-298 slow search on large workspaces. Investigation running.Start with GitHub and your support inbox. Add Sentry, Linear and Slack when you are ready.