Skip to content
Critique/docs
Platform

Dashboard

Navigate Critique's workspace hub — overview, automation, review runs, usage, and settings.

The dashboard at /dashboard is the control center after you connect GitHub. Everything below assumes you are signed in.

SectionPurpose
OverviewRecent review activity, connection status, quick links
AutomationInstallations, repo list, review policy, secrets
CheckpointPre-review trust gates across repos
Review runsHistory, live progress, rerun and Remedy actions
BuilderCloud builder jobs (when enabled for your account)
UsageCredits, token-style usage breakdown, model spend
SystemOperator-oriented status (when visible for your role)
HelpSupport paths and product links
ReferralsInvite links and bonus credit rewards

Chat settings (under Settings in the product nav) cover default chat models and optional API keys — separate from per-repo review policy.

Automation

Use Automation to answer: “Which repos are reviewed, with what rules?”

  • Sync GitHub App installations and see which repositories are active.
  • Edit installation-wide policy defaults.
  • Override policy per repository when one service needs stricter security or looser merge blocking.
  • Manage repository secrets for sandbox and validation runs that need credentials.

Policy fields are documented in Policy fields.

Review runs

Review runs is the audit trail for automated and manual reviews:

  • Filter and open any run by repository or verdict.
  • Live view shows in-progress stages when a run is still executing.
  • From a completed run you can copy a fix prompt, start Remedy, queue BYOA agents (Cursor, Claude, Codex), or open the matching PR on GitHub.

Public URLs embedded in GitHub comments (/review-run/...) resolve to the same run after sign-in.

Checkpoint

Dashboard → Checkpoint lists gate events and links into per-repo configuration at /dashboard/[owner]/[repo]/checkpoint. See Checkpoint.

Usage

Usage shows your credit pool for the current period, consumption by surface (review, chat, Remedy, embeddings), and enough detail to explain a spike. It is the source of truth when Billing & credits mentions “check the dashboard.”

Account and settings

  • Account — Profile and password.
  • Settings — Chat preferences; OpenRouter BYOK; BYOA API keys (Cursor, Anthropic, OpenAI); browser-local Anthropic key for Claude Code chat handoff.
  • Referrals — Share links; successful referrals add bonus credits to your balance.

Typical workflows

Onboard a new repo
Automation → confirm repo appears → set or inherit policy → open a test PR → confirm check + comment → adjust strictness if merge blocking is too aggressive.

Investigate a failed check
Review runs → open run → read findings → use inline links on GitHub or Fix prompt / Remedy from the run page.

Control spend
Usage → identify heavy models or large PRs → raise strictness only where needed → consider Standard vs Pro vs Ultra plan on Pricing → optional BYOK or BYOA in Settings.

Fix findings externally
Review runs → Queue Cursor / Claude / Codex agent (keys in Settings) or Fix prompt → paste into Copilot / Claude Code. See BYOA.