Seer
Learn about Seer, Sentry's AI agent, and managing generative AI features.
Seer is Sentry's AI debugging agent. It uses Sentry's rich context (issue details, tracing data, logs, and profiles) to help you troubleshoot and fix errors and performance issues faster.
Seer provides end-to-end debugging tools to help you troubleshoot and fix errors and performance issues faster, often before they are even merged:
- AI Code Review: Have Seer review your code changes in GitHub, catching bugs before merging pull requests.
- Root Cause Analysis: Automatically scans issues as they come into Sentry, feeding additional context to your alerts and automating triage
- PR Creation: A collaborative agent, skilled at root cause analysis and solving issues
- Coding Agents: Delegate Seer's analysis to an external coding agent for further debugging and fixes
To start using Seer's capabilities in your Sentry-enabled projects, follow these steps:
Connect to GitHub through the Sentry GitHub integration. You can follow the steps in Seer settings to get started. Note: Seer can only be integrated with the cloud version of GitHub
Select which repositories you want Seer to have access to
Select which features to turn on - AI Code Review, Root Cause Analysis, PR generation of fixes
Seer is an add-on to your Sentry subscription. By enabling it, you are signing up for active contributor pricing for this feature. Any person who creates 2 or more PRs in a month in a Seer-enabled project will be billed. You can learn more about Seer pricing here.
AI Code Review helps you reviews your code changes, predicting errors and offering suggestions for improvement before merging pull requests. This feature is only available on GitHub.
Seer is able to automatically analyze issues as they are ingested by Sentry by combining all of the relevant context from your code with Sentry's telemetry data to provide a best-in-class issue debugging experience. This initial review is used to:
- Determine if the issue can be automatically analyzed and fixed using Seer's Root Cause Analysis
- Augment issue alerts with a helpful analysis about the problem
- Add an initial guess for what the potential problem is to the issue details page
Once the Root Cause Analysis has run, Seer will provide remediatiion recommendations.
You can prompt Seer to generate PRs to fix your issue, and push it to GitHub. You can add more context in natural language, or any external supporting information, to help Seer generate a better PR. Note: You must install the Seer GitHub app to use this feature.
You can delegate Seer's root cause analysis to an external coding agent, such as Cursor, for further debugging and fixes.
Seer is a powerful debugging agent, with access to a wide variety of data sources and tools. While debugging issues, it may examine:
- Issue Context: Error messages, stack traces, and event metadata from your Issues
- Tracing Data: Distributed traces and span information
- Logs: Structured Logs from your application (beta)
- Your Codebase: Relevant code from linked GitHub repositories, with support for multiple repos for distributed services
- Performance Data: Profiles and performance metrics
- Interactive Feedback: Your input and guidance during the process
Sentry includes strong guarantees for privacy and security of your data. At a glance:
- Sentry does not train generative AI models using your data by default and without your permission.
- AI-generated output is shown only to you and other authorized users in your account.
You can learn more about our data privacy practices in the security and legal docs.
If you would not like to show generative AI Features at all in your Sentry account, go to organization settings and turn off the
Show Generative AI Featurestoggle.If you want to use some generative AI features, while disabling others, you can bulk manage these settings by going to Seer settings and selecting from the dropdown at the top of the table in each tab.
For organizations that need to prevent code generation, or adding additional context in Alerts, you can find these configurations under Advanced Settings in Seer settings.
Note: Disabling code generation prevents Sentry from automatically generating code, or pushing code to your codebase. This setting does not impact AI chat sessions, where it is possible to prompt AI chat to emit code snippets and examples. None of this affects workflows involving your own coding agent. Learn more about code generation.
Our documentation is open source and available on GitHub. Your contributions are welcome, whether fixing a typo (drat!) or suggesting an update ("yeah, this would be better").