User guide
Each guide is one task: a short clip, then a few steps with a screenshot each. No API, no configuration file.
To see how a feature works, or to automate it, follow the “Go further” link at the end of each guide.
The screenshots are recorded from the dashboard by a script. They are recorded again when the dashboard changes, so they match what you see.
Developer
- Read the review of a pull request: Open the report of one pull request: its findings by severity, and each file that was reviewed.
- Dismiss a false positive, and say why: Mark a finding that is not a real problem, so the review stops raising it and the rule learns from your reason.
- Find the findings of one pull request: Filter the Findings page down to one pull request.
- Read the rule behind a finding: See which rule raised a finding, what it says, and where to change it.
- Ask the support assistant why a check failed: Ask the assistant a question from any page; it answers from your findings and the docs.
- Give your coding agent or a script access: Create an access token that lets a coding agent or a script act on your behalf.
- Follow a review while it runs: Watch a pull request's review run, step by step, on the Jobs page.
Org admin
- Open your product's overview and its report: See a product's repositories, their open findings, and one report across all of them.
- Create a product and add its repositories: Group the repositories of one service into a product, so its teams and goals cover them all.
- Scan the default branch of a repository: Review the code already on the default branch, not only what pull requests change.
- Turn on a pack and pick its rules: Choose which kinds of files are reviewed, and which of a pack's rules run.
- Turn an ADR into a rule: Let OpenTremor read your repository's decision records and draft a rule from each.
- Set a product goal and follow it: Say where a product is heading, in measured steps, and see which pull requests move it.
- Write a rule of your own: Add a rule in plain words; the next review checks it.
- Find the rules that misfire: See how often reviewers agree with each rule, and which rule changes the evidence supports.
- Share a report with someone outside the org: Get a link to a report that anyone can open, without an account, until it expires.
- Make a report template: Change how reports look, and which layout is used by default.
- Choose which findings get a code suggestion: Let reviews post fixes on pull requests, for the findings that matter.
- Invite a colleague and give them a role: Create an invitation for one person, with the role they get when they join.
- Create a team and route its repositories to it: Group people into a team and give it the repositories it reviews.
- Get alerts in Slack, by email or on a webhook: Send a message to a channel when a pull request check is blocked.
- Create a role with only the rights it needs: Make a role from a chosen set of permissions.
- Create a service account for CI: Create an API key with its own role, for a CI pipeline or an MCP client.
- Turn on single sign-on for the org: Let members sign in through your identity provider.
- Find who changed what: Search the organization's audit log, and see what each change did.
- Connect GitLab: Connect your self-managed GitLab, so merge requests are reviewed.
- Send findings to Jira: Open Jira issues for the findings worth tracking, kept in step with each finding's status.
- Connect GitHub and choose repositories: Create your organization's GitHub App and install it on the repositories to review.
Org owner
- Cap the LLM spend and choose a model: Set a monthly budget for the model, and the model each product uses.
- See what the org uses and spends: See this month's model spend, and how much the organization reviews.
- Bring your own LLM connection: Add a connection to a hosted model or to your own endpoint.
Platform admin
- Find an organization and change its plan: Open any organization on the deployment and change its plan.
- Choose which models the orgs can use, and their price: Keep the catalog of allowed models, with the prices that count against each org's budget.
- See usage across the platform: See this month's model spend for the whole deployment, and per organization.
- Restrict access by IP address: Say which addresses may reach the API, the dashboard and the docs, path by path.
- Set up backups: Choose when backups run and where they go, and run one now.
- Stream the audit log to your SIEM: Send every audit event of the deployment to Splunk, Elastic, Datadog or a syslog server.
- Read the platform audit log: Search every privileged action on the deployment, in every organization.
- Let orgs reach your GitLab or LLM endpoints: List the hosts on your network that organizations may connect to.
- Turn on the support assistant: Offer the assistant to the organizations of the deployment.
- Turn agent access on or off: Decide whether people can connect agents and create access tokens.
- Check migrations after an upgrade: See which data migrations ran after an upgrade, and what they changed.
- Find a user and deactivate them: Find any account on the deployment and stop it from signing in.
- Lock an organization: Stop an organization's members from using it, without deleting anything.
- Decide who can sign up and how people sign in: Turn self sign-up off, and choose the sign-in providers.