Sentry
Errors and crash triage
Sentry lets agents search recent errors, inspect stack traces and affected releases, and summarize what broke after a deploy. Each user signs in to their own Sentry account through the browser.
What an agent can do with Sentry
Find new errors
The agent searches grouped issues with search_issues and error events with search_errors. It can also search logs, traces, metrics, profiles and session replays. Each search tool covers one dataset, so a log search does not return spans.
Explain a crash
The agent reads an issue with get_issue_details and the full stack trace with get_event_stacktrace. It can read breadcrumbs and tag values too. Then it can ask Seer for a root cause analysis and a code fix with analyze_issue_with_seer.
Check a release
The agent finds recent releases with find_releases and reads one version with get_release_details. Use this to see which version went out last before you look at new errors after a deploy.
Triage and assign issues
With the Triage Issues skill, the agent resolves, reopens, assigns or ignores an issue with update_issue. It adds a comment to the issue activity with add_issue_note. It can also link a Jira ticket or a GitHub pull request to the issue.
How to connect Sentry to OpenBot
Open Sentry in OpenBot
Press Open in OpenBot on this page, or find Sentry in the app's Apps tab. Get the app first if it is not on this computer.
Sign in
Press Connect and finish the sign-in in your browser. OpenBot keeps the token on the computer that runs OpenBot. You can disconnect at any time.
Give an agent a task
Ask an agent for work that needs Sentry, such as one of the example prompts below. The agent uses the plugin's tools to do it.
Example prompts3
SentryWhat new errors appeared since yesterday's deploy?
SentryExplain the top crash in the mobile project and its likely cause.
SentryHow healthy is the current release compared to the last one?
Apps1
Sentry
Error search, issue inspection, and release health, over Sentry's MCP server with browser sign-in.
Information6
- Developer
- sentry.io
- Category
- Coding
- Version
- 1.0.0
- Website
- sentry.io
- Privacy policy
- sentry.io/privacy
- Terms
- sentry.io/terms
Questions
Is the Sentry MCP server official?
Yes. Sentry builds and hosts the server. Its source is in Sentry's getsentry/toolkit repository on GitHub, next to the Sentry CLI. Sentry says that the server is mainly for coding agents that work with a person in the loop.
Does the Sentry plugin need an API key?
No. Each user signs in to their own Sentry account through the browser. OpenBot does the sign-in itself, keeps the tokens in encrypted storage, and adds the Authorization header. The plugin catalog holds no secret.
What can an AI agent do with Sentry?
Sentry puts its MCP tools in skills: Inspect Issues & Events, Seer, Documentation, Triage Issues, and Manage Projects & Teams. With them, an agent searches errors and logs, reads traces and replays, gets a Seer root cause, updates issues, and creates projects, teams, DSNs and monitors.
What data does Sentry get?
Your workspaces, conversations and files stay on the computer that runs OpenBot. Sentry gets the requests that the agent sends, such as a search query or an issue update. The tools work on the organizations and projects that your Sentry account can access.
Sources
Every statement about Sentry comes from the pages below, checked on . They can change after that date.