Configuration Options You Should Know
Get started with the getting started guide if you haven’t set up Greptile yet.Dashboard Configuration
Learn more about all available settings in the key features guide.Adjust how many comments Greptile makes
The severity threshold is your primary tool for controlling how many comments Greptile makes:Low Threshold
- Comments on potential issues
- Includes minor suggestions
- Good for teams that value detailed feedback
Medium Threshold
- Mix of minor and major issues
- Filters out most nitpicks
- Recommended starting point
High Threshold
- Fewer, high-impact comments
- Focuses on bugs and security
- Good for teams that find the bot too noisy
Adjust what types of comments Greptile makes
You can control which types of issues Greptile comments on:- Syntax: Basic code structure and formatting
- Logic: Potential bugs and logical issues
- Style: Coding conventions and best practices
Help Greptile Learn About Your Codebase and Patterns
Learn more about how Greptile learns in the memory and learning guide.Effective Comment Interaction
React to Comments
Explain Your Feedback
@greptileai Actually, we prefer list comprehensions because they're more readable
Ask Questions
@greptileai
to ask for clarification: @greptileai Can you explain why this is a security issue?
Be Consistent
Learning timeline
- Filtering suggestions to what your team values
- Understanding your codebase context
- Recognizing your team’s coding patterns
What does Greptile learn from?
- Your reactions to its comments
- Other developers’ comments on PRs
- Which suggestions get implemented vs ignored
- Team discussions about code patterns
Managing Custom Context
Types of Custom Context
See the full custom context guide for detailed instructions.Rules
- Enforce coding standards
- Specify preferred patterns
- Set architectural guidelines
Files
- Style guides
- Architecture docs
- API standards
Other
- Project context
- Domain knowledge
- Special requirements
Effective Context Rules
Context File Best Practices
Point Greptile to your existing documentation:- Style guides: Language-specific coding standards
- Architecture docs: System design and patterns
- Security policies: Compliance and security requirements
- API standards: Consistent API design patterns
- Testing strategies: How your team approaches testing
Advanced Configuration
Repository-Level Settings
Usegreptile.json
in your repository root for per-repo configuration. Learn more in the greptile.json guide:
Analytics Monitoring
Track these key metrics in your dashboard:- Average time to merge: Are reviews helping or slowing down merges?
- Comment engagement: High upvote rates indicate valuable feedback
- Addressed comments per MR: Shows team adoption and value
Integration Optimization
Trigger Configuration
Trigger Configuration
- Specific authors (new team members)
- Branch patterns (feature branches)
- PR labels (needs-review)
- File types (critical system files)
Pattern Repositories
Pattern Repositories
- Add frontend repo to backend reviews
- Include shared library repositories
- Cross-reference microservice dependencies
Team Adoption Strategies
For New Teams
Start Conservative
Encourage Engagement
Iterate Based on Feedback
Add Context Gradually
Getting Support
Documentation
Community Support
Common Pitfalls to Avoid
Troubleshooting & FAQ
Common Issues
Greptile isn't reviewing my PR
Greptile isn't reviewing my PR
- Your PR doesn’t match configured trigger rules (labels, authors, branches)
- Changes are in excluded files or directories
- Repository indexing is still in progress (check for “updating” tag in dashboard)
- Try manual trigger: comment
@greptileai
on the PR (see manual triggers) - Check your team’s trigger configuration in dashboard settings
- Verify the repository is properly connected and indexed
Getting too many irrelevant comments
Getting too many irrelevant comments
- Increase severity threshold in dashboard settings (move from Low → Medium → High)
- Disable comment types you don’t need (syntax if you use a linter)
- Downvote irrelevant comments consistently to help Greptile learn
- Add custom rules to clarify your team’s preferences (see custom context)
Comments aren't helpful or accurate
Comments aren't helpful or accurate
Repository indexing is taking too long
Repository indexing is taking too long
- Large repositories (100k+ lines): Several hours for initial index
- Subsequent updates: Much faster, only processes new changes
- You’ll see “updating” tag in dashboard while indexing is in progress
Greptile stopped working after recent changes
Greptile stopped working after recent changes
- Repository permissions or access tokens may have expired
- Trigger rules may have been updated to exclude your PRs
- Repository may have been disconnected accidentally
- Check repository connection status in dashboard
- Verify trigger rules still match your workflow
- Try manual trigger to test if Greptile responds (see manual triggers)
Frequently Asked Questions
How long does it take for Greptile to learn our preferences?
How long does it take for Greptile to learn our preferences?
- Week 1: High noise as Greptile learns basic patterns
- Week 2-3: Significant improvement in comment relevance
- Month 2+: Highly tuned to your team’s specific preferences
Can I use Greptile on old PRs or closed PRs?
Can I use Greptile on old PRs or closed PRs?
@greptileai
(see manual triggers). This works on:- Old PRs from before Greptile was enabled
- Closed or merged PRs (for learning purposes)
- Draft PRs that don’t normally trigger automatic reviews
How does Greptile handle large PRs or complex changes?
How does Greptile handle large PRs or complex changes?
- Summary of overall changes in natural language
- Sequence diagrams for complex changes
- Lower confidence scores for large, risky changes
- Context-aware comments that understand cross-file dependencies
What happens if I disagree with a Greptile comment?
What happens if I disagree with a Greptile comment?
- Downvote the comment so Greptile learns
- Reply explaining why:
@greptileai This pattern is actually preferred because...
- Ask for alternatives:
@greptileai What's another way to handle this?
- Greptile will learn from your feedback and adjust future comments
Can I configure different settings for different repositories?
Can I configure different settings for different repositories?
- Dashboard settings apply to all repositories in your team
greptile.json
files in repository roots can override dashboard settings (see greptile.json guide)- Custom context can be scoped to specific repositories or file types (see custom context)
- Different teams in your organization can have completely different configurations
How do I know if my custom context is being used?
How do I know if my custom context is being used?
- “Last Applied” column shows when each context item was used
- Click on any context item to see usage statistics
- “Recent Usage” section shows specific comments where the context was applied
What should I do if Greptile suggests a security issue?
What should I do if Greptile suggests a security issue?
- Even if you disagree, consider the suggestion carefully
- Ask for clarification:
@greptileai Can you explain the security risk here?
- Research the potential vulnerability independently
- If it’s a false positive, downvote and explain why to help Greptile learn
@greptileai Actually, this pattern is preferred in our codebase because...