cubic
Also known as: cubic, cubic.dev
AI code review agent that reviews every GitHub pull request with whole codebase context, catches bugs, enforces plain English rules, and runs background scans that find and fix issues.
cubic is an AI code review platform built for the moment when AI generated code has made review the bottleneck. When a pull request opens, cubic reviews the changes automatically in a secure isolated container, with context into the whole codebase rather than just the diff, and comments inline on the lines that matter. Its reviewer navigates a repository the way a developer would, jumping to definitions and grepping through code, and its analysis is repository wide, so it catches issues that span multiple files. A Y Combinator backed startup used by teams at Cal.com, n8n, and the Linux Foundation, cubic reports helping teams merge pull requests roughly a quarter faster.
Beyond flagging bugs, cubic learns each team's conventions from its codebase and its senior engineers' past review comments, so feedback grows more relevant and less noisy over time, and teams can write rules in plain English to enforce standards without hand writing linter configs. It generates pull request descriptions that explain the change and its impact, and it offers fixes: simple ones commit in a single click, and harder ones are handled by background agents. On higher tiers, cubic runs scheduled background scans in which large numbers of agents comb the whole repository for bugs and security issues, triage the findings, notify owners, create tickets, and open one click fix pull requests.
A distinctive part of cubic is that the platform is designed around the human reviewer, taking cues from tools like Linear and Superhuman. It ships a desktop app, orders diffs logically instead of alphabetically, groups related changes, draws diagrams of architecture changes, and supports a pull request inbox and stacked pull requests. It integrates two way with GitHub and the IDE and pulls context from Jira, Linear, Asana, Notion, and Confluence, so a change that references a ticket is reviewed with that context. cubic is SOC 2 compliant and reviews code without storing it or using it to train models.
cubic is free for public repositories and offers a genuinely usable free tier with a capped number of monthly reviews, custom agents, automatic descriptions, and project tool integrations. Paid plans are per developer: a Team plan in the range of thirty to forty dollars a month adds a large monthly reviewed line allowance and custom agents, and a Pro plan around eighty to one hundred dollars a month adds the nightly codebase scans, deeper integrations, and automatic fix pull requests. Enterprise adds custom agreements. Reviewers note the standout background scanning is the premium draw and was still maturing in early 2026, so smaller teams should trial before paying up.
Vendor details
Canonical URL
https://www.cubic.dev
Category
Coding agent
Subcategory
AI code review agent
Funding status
Independent. A Y Combinator backed startup founded by engineers who hit the code review bottleneck, cubic is used by teams at Cal.com, n8n, the Linux Foundation, and PostHog. Specific funding figures are not disclosed in available sources.
Company status
independent
Use cases & customers
Primary use cases
Target customers
Deployment options
Integrations
Integrates two way with GitHub via a two click app and works in the IDE, and pulls project context from Jira, Linear, Asana, Notion, and Confluence so a change that references a ticket is reviewed with that context. It acts by leaving inline comments, committing one click fixes, creating tickets, and opening fix pull requests. Language agnostic across popular languages.
In practice
Your team is drowning in pull requests and rubber stamping them. cubic reviews each PR against your whole codebase, flags real bugs on the exact lines, and enforces the rules you wrote in plain English.
You maintain a busy open source project alone. cubic is free for public repos, learns your conventions, and reviews contributor PRs so you spend less time on routine feedback.
You worry about latent bugs already merged. On higher tiers cubic runs nightly scans where many agents comb the repo for bugs and vulnerabilities, then file tickets and open fix pull requests.
Sources & related URLs
Related / legacy domains
Capability coverage
6.5 / 14 capabilities · 46%
| Integrations & Tool CallingIntegrates two way with GitHub, works in the IDE, and pulls context from Jira, Linear, Asana, Notion, and Confluence, and acts by leaving inline comments, committing one click fixes, creating tickets, and opening fix pull requests, broad integration with real action. | Full |
|---|---|
| Workflow OrchestrationOrchestrates the review pipeline (review, comment, suggest and commit fixes) and, on higher tiers, runs a background scan where many agents triage findings, create tickets, and open fix PRs, real orchestration, though the multi agent scanning is a premium and still maturing capability rather than the core loop. | Partial |
| Knowledge Grounding & RAGGrounds each review in repository wide context, navigating the codebase the way a developer would with grep and jump to definition, learns the team's patterns from history, and pulls context from Jira, Linear, Notion, and Confluence to understand why a change was made, a comprehensive headline grounding capability. | Full |
| Human Oversight & GuardrailsAugments human review rather than replacing it: humans still review and merge, cubic enforces plain English custom rules on every pull request, and offers one click fixes the developer chooses to commit, a strong human oversight surface, though it advises and flags rather than hard blocking a merge. | Partial |
| Security, Identity & GovernanceIs SOC 2 compliant, reviews code in a secure isolated container, never stores code or uses it for training, offers data ownership, and provides enterprise MSA and DPA agreements, a solid security posture, though a full identity and governance matrix (SSO, RBAC, audit) is not enumerated. | Partial |
| Observability & AuditabilityProvides review visibility through an organized pull request inbox, logically ordered diffs, architecture change diagrams, and codebase chat, and its background scans triage findings and notify owners, real observability into code changes and issues, though centered on review rather than agent execution tracing. | Partial |
| Memory & State PersistenceSelf learning adapts to team conventions and remembers maintainer corrections over time, but this is feedback loop adaptation grounded in the codebase and review history, counted under knowledge grounding, not a distinct runtime agent memory product. | Unable to verify |
| Deployment & Data ResidencyDelivered as a hosted SaaS via a GitHub app; reviews run in an isolated secure container, but no self host, on premises, or data residency deployment option is documented. | Unable to verify |
| Prebuilt Agents, Templates & PacksSupports user defined custom agents (up to five to ten per tier) and plain English custom rules that act as reusable review building blocks, custom agent scaffolding, though not a library of prebuilt agents or templates shipped by the vendor. | Partial |
| Triggers & Channel CoverageAutomatically triggers on pull request open and runs scheduled background scans, and works across GitHub, the IDE, a desktop app, and the web platform, real trigger and channel coverage, though scoped to the review workflow. | Partial |
| Model Flexibility & RoutingUses AI models internally to review and scan code, but no user facing model choice, bring your own key, or routing gateway is documented. | Unable to verify |
| APIs, SDKs & MCP ExtensibilityCustom agents and plain English rules extend behavior within the product, but no public API, SDK, or MCP server for building on the platform is documented. | Unable to verify |
| Testing, Debugging & OptimizationBug and vulnerability detection is the core product: it reviews every pull request for bugs and quality issues, enforces custom rules, and runs background scans with many agents to find serious bugs and security issues across the codebase, then generates tested fix pull requests, a headline testing, debugging, and quality capability. | Full |
| Browser & Computer UseReviews and scans code in an isolated secure container, navigating the codebase with developer tools like grep and jump to definition and generating tested fixes, a form of sandbox computer use, though not general browser automation. | Partial |
Pricing
Free (20 reviews/mo; free for public repos) · Team ~$30-40/dev/mo (40k lines) · Pro ~$79-99/dev/mo (nightly scans, auto-fix) · Enterprise custom
Per-developer/month subscription tiers with a monthly reviewed-line allowance (added/deleted diff lines cubic reads); free tier capped by monthly review count; enterprise custom
Included quota
Free/Starter ($0): 20 PR reviews/mo, up to 5 custom agents, auto PR descriptions, custom context, Jira/Linear/Asana/Notion integrations, unlimited AI wikis; free & unlimited for public repos. Team ~$30-40/dev/mo: AI PR reviews, ~40,000 reviewed lines/mo, up to 5 custom agents, Jira/Linear/Asana. Pro ~$79-99/dev/mo: adds nightly codebase scans, Ultrareview, up to 10 custom agents, Confluence integration, auto-fix PRs, deeper integrations. Enterprise: custom (MSA/DPA). Reviewed lines = added/deleted diff lines cubic reads (generated/vendored/binary/ignored files excluded); resets each billing period; incremental reviews count only new changes. SOC 2; code not stored or used for training.
What is public
Free tier and tier structure are public; exact Team ($30 vs $40) and Pro ($79 vs $99) prices vary across sources. Enterprise is custom.
Billing mechanics
Per-developer/month tiers with monthly reviewed-line allowances; free tier capped by review count; free & unlimited for public repos.
Cost watchouts
Reviewed-line allowances can be exceeded by high PR volume/large diffs; Pro roughly doubles+ the per-seat cost; free tier caps at 20 reviews/mo.
Variable cost rationale
Per-developer pricing is predictable, but each tier caps reviewed lines per month (e.g. ~40,000 on Team) and the free tier caps monthly review count, so high PR volume or large diffs can push a team to a higher tier. The main cost step is moving to Pro (~$79-99) for nightly scans and auto-fix. Exposure is moderate: bounded by per-seat tiers but sensitive to review volume and team size.
Additional watchouts
Nightly background scans (the standout Pro feature) were still maturing/early-access in early 2026. Free tier review cap (20/mo) hits quickly for active teams. Pro (~$79-99/dev) can be a stretch for small teams. GitHub-centric. Compliance-strict teams should verify data handling.
Overage / add-ons
Usage is metered by reviewed lines (added/deleted diff lines cubic reads) against the monthly allowance; the free tier is capped by monthly PR-review count. Excess beyond allowance requires a higher tier.
Sales call required
No — self-serve available
Free / trial
Free/Starter plan ($0): 20 PR reviews/month, up to 5 custom agents, auto PR descriptions, custom context, integrations (Jira/Linear/Asana/Notion), unlimited AI wikis. Free and unlimited for public repositories. 2-week free trial (no credit card).
Lowest paid plan
Team at ~$30-40/developer/month: AI PR reviews, ~40,000 reviewed lines/month, up to 5 custom agents, and Jira/Linear/Asana integration. Sources cite $30 and $40 across early-to-mid 2026.
Commercial notes
Priced above CodeRabbit; positioned on review depth (repo-wide context, custom rules, background scans). Free tier genuinely usable (esp. open source). Pro's nightly scans are the premium draw. Value highest where review quality is inconsistent or seniors are bottlenecked.
Key ambiguities
Exact Team and Pro prices differ across sources; the value of the premium nightly-scan feature depends on its maturity for your codebase.
Cancellation / refund
Free tier and 2-week trial (no credit card). Monthly per-developer tiers. Enterprise via MSA/DPA. Specific cancellation/refund terms not detailed.
Support SLA / resale
Standard support on paid tiers; Enterprise adds custom MSA/DPA. SOC 2 compliant; code reviewed in a secure container, never stored or used for training. No reseller/whitelabel program surfaced.
Missing data
Exact Team/Pro prices vary by source ($30-$40 / $79-$99); per-reviewed-line economics beyond allowances and enterprise pricing not fully public.
Related vendors
- Cognition — Maker of Devin, an autonomous AI software engineer
- Aider — Open-source, model-agnostic terminal coding agent that works…
- Anthropic Claude Code — Anthropic's terminal and IDE-native coding agent with sub-agent…
- Augment Code — AI coding assistant evolving into an AI-native engineering platform
- Blackbox AI — AI coding assistant for code generation and developer workflows
- Blink — Agentic AI app builder that turns a natural language prompt into a…