Pricing
Pay for the engine running in the right places.
Free CLI is the local engine — maps, algorithmic judgment, pre-change gates, evidence packets verified offline. Teams pay to run it on every PR, in a shared dashboard, or in a scoped private-deployment pilot when hosted services are policy-blocked.
The buying path
Most teams take this in order. Skip rungs as needed.
-
1
Install the free CLI.
pip install roam-code && roam init— a 100k-LOC repo indexes in seconds. Apache 2.0, local, no telemetry. Try before you buy. - 2 Commission a PR Replay audit. Paid one-shot engagement (Team $2,500 / Deep $6,000) against your last 30 or 90 PRs with a written report. See tiers. 50% credits toward Roam Review within 60 days.
- 3 Join Roam Review early access. Planned hosted GitHub App posting a structural verdict on every PR — flat tiers from $99/mo, no per-seat math. Tier matrix below.
- 4 Add Roam Cloud once you want history. Dashboard layer behind Review — health, debt, danger-zone trends. From $19/repo/mo. See plans.
- 5 Discuss private deployment if your stack requires it. Scoped private-deployment pilot when hosted Review or Cloud is policy-blocked. Talk to us.
Want a paid evaluation before subscribing? Commission a PR Replay — Team $2,500 / Deep $6,000, written report on your last 30 or 90 PRs, 50% of the fee credits toward Roam Review.
| Free CLI | Roam Review | Roam Cloud | Self-Hosted | |
|---|---|---|---|---|
| Price | Free, forever | from $99 /mo | from $19 /repo/mo | Custom |
| Billing | — | monthly or annual | monthly or annual | annual |
| License | Apache 2.0 | commercial | commercial | commercial |
| 241 CLI commands | Yes | Yes | Yes | Yes |
| 227 MCP tools for AI agents | Yes | Yes | Yes | Yes |
| 28 languages, 6 cross-language bridges | Yes | Yes | Yes | Yes |
| Runs entirely on your machine | Yes | PR diffs in our cloud | metrics in our cloud | Yes (your network) |
| Posts a verdict on every pull request | — | Yes | — | Yes (incl. Review) |
| Dashboard with metrics over time | — | via Cloud | Yes | Yes (incl. Cloud) |
| Tamper-evident audit-trail file (in-toto v1) | Yes (local file) | Yes (per PR) | browser only | Yes (full pipeline) |
| SARIF export for GitHub Code Scanning | Yes | Yes | Yes | Yes |
| SSO/SAML, custom rules, audit logs | — | SSO on Business+ | SSO on Growth | Yes |
| Framework control mapping (maps to, not certified) | via Governance Pack | via Governance Pack | via Governance Pack | Planned |
| SLA + dedicated support | community only | business hours | business hours | 24/7 + named contact |
| Free for open-source projects | Always | Yes, forever | — | — |
| First-charge refund | — | 14-day EU withdrawal | 30 days when available | via order form |
Pricing reflects rates as of ; will be locked at the rate you sign at for the term you committed to. See refund policy.
Trying it solo? Free CLI. Team of 5-150? Roam Review. Multi-team org? Roam Cloud. Regulated or private-network requirement? Talk to us.
The CLI is the same engine in all three. The difference: the CLI runs when an engineer remembers to pipe a diff. Roam Review runs on every PR. Roam Cloud collects the metrics. Private-deployment pilots scope what must run inside your network.
Product tiers
Roam Review Early access
A structural and algorithmic second opinion on every pull request — runs alongside your semantic reviewer, not instead.
- Structural Risk Review — clones-not-edited, layer violations, blast-radius spikes, hidden coupling
- Algorithmic Risk Review — accidental O(n²), N+1 queries, regex-in-loop, JSON-parse-in-loop, quadratic strings
- Test Reachability Review — which tests actually reach the changed lines
- Refactor Safety Review — simulation of moves and renames before merge
- Audit-grade evidence — signed in-toto attestations attached to every PR
Roam Cloud Early access
A dashboard for your code health over time. We never see your source.
Roam Self-Hosted
Private-deployment pilots for teams that cannot use hosted Review or Cloud.
Limits and definitions
Roam Review tiers are flat-rate with explicit usage caps. We warn at 80% and 100% of every cap; we never apply overage charges unless you opt in. Below the caps, your bill never moves.
| Tier | Repos | Active PR authors | Reviews / month | Overage (opt-in) |
|---|---|---|---|---|
| Starter $99/mo | 3 | 5 | 100 | $12/extra author/mo to 20 · $0.50/extra review |
| Team $299/mo | 20 | 30 | 900 | $10/extra author/mo to 75 · $0.40/extra review |
| Business $799/mo | 100 | 100 | 3,000 | $8/extra author/mo to 150 · $0.30/extra review |
| Scale $1,499/mo · annual | negotiated | 250 | 8,000 | $6/extra author/mo · custom quote above 300 authors |
Definitions
- Active PR author — a unique human author of a reviewed pull request in the trailing 30 days. Bots, Dependabot, Renovate, and generated release-PR authors do not count.
- Review — one Roam structural-risk verdict posted on a pull request (open or push). Re-runs on the same PR within 24h count as one review.
- Repo — one git remote you have installed the Roam Review GitHub App on. Monorepos count as one repo unless you ask for split billing.
- Warnings — at 80% of any cap we email the org admin; at 100% we surface a banner in the dashboard and offer either upgrade or opt-in overage.
- No surprise bills — overage rates above are opt-in. Default behaviour at the cap is "next review skipped, banner shown" until you upgrade or enable overage.
- Renewal protection — launch customers keep their package through their next renewal. Renewal increases are capped at 15% unless you have exceeded a cap for two consecutive months.
Why flat tiers, not per-developer pricing?
Per-developer pricing forces a procurement comparison with semantic reviewers and creates a per-seat expansion trap. Roam reviews architecture, not semantics — different layer, complementary to whatever code-review bot you already run. Flat tiers with usage caps keep the buying decision at engineering-manager discretion and let you pilot without a per-seat headcount conversation. We will revisit usage-based pricing once we have measured COGS across at least 50 paying Review customers.
Want to try a paid engagement before subscribing?
PR Replay is a one-shot paid audit (Team $2,500 / Deep $6,000) that runs Roam against your last 30 or 90 PRs and produces a written structural-review report plus a founder walk-through. 50% of the fee credits toward a Roam Review subscription if you start within 60 days, so the engagement is close to free if you go on to subscribe. See /audit for the deliverable spec, FAQ, and credit math.
Pricing FAQ
What if my team doesn't use AI agents yet?
The free CLI is a standalone code-intelligence layer — health scores, dead-export detection, blast-radius analysis, SARIF export for GitHub Code Scanning, in-toto v1 audit-trail files. None of it assumes an agent. Roam Review's PR verdict is just as useful for human-authored PRs as for agent-authored ones; the structural and algorithmic risk classes do not care who wrote the diff.
What if we already use CodeRabbit, Greptile, or another semantic reviewer?
Keep them. Roam Review is a different layer — semantic reviewers comment on diff readability and style; Roam reviews structure (clones-not-edited, layer violations, blast-radius spikes) and algorithmic risk (accidental O(n²), N+1, regex in loop). Both are flat tiers, both run on every PR, both ship their own verdict. We routinely sit alongside a semantic reviewer in customer pipelines.
Can I trial Roam Review without a credit card?
The CLI is Apache 2.0 and free forever — install with
pip install roam-code and run the same engine
locally. Roam Review (the planned GitHub App) is expected to include a
14-day trial on paid tiers at launch. For a paid pre-subscription
evaluation across a real PR history, see
PR Replay — 50% of the fee credits toward
a Review subscription if you start within 60 days. We expect public
open-source repositories to get Roam Review free at launch; email
hello@roam-code.com
with a repo link.
How do Roam Review and Roam Cloud fit together?
Roam Review posts a verdict on every PR; Roam Cloud is the
planned dashboard that aggregates those verdicts over time —
health, debt, danger-zone trends per repo and per author. The
current public Cloud pricing target is Starter at $19/repo/mo,
Team at $99/mo for 10 repos, and Growth at $299/mo. You can run
Review without Cloud (verdicts ship to the PR), Cloud without
Review (it ingests local roam health runs), or both
together once you want history.
What happens if I exceed the cap mid-month?
Soft cap by default. We email the org admin at 80% of any cap and surface a banner at 100%. The default at 100% is "next review skipped, banner shown" — Roam Review keeps the queue rather than dropping PRs, and your bill never moves unless you opt in to overage. You can either upgrade tiers (pro-rated) or enable the per-tier overage rates listed in the limits table. No surprise bills.
Do you offer OSS or educational discounts?
Roam Review is free, forever, on public repositories — see the comparison table above. We do not gate the CLI's features by license, so educational users can run the full engine locally at zero cost. For an institutional Review or Cloud rollout (a department, a course, a research group), email hello@roam-code.com with a description and a repo link.
Can I self-host?
Roam Self-Hosted is a private-deployment pilot path, scoped by SOW for teams that cannot use hosted Review or Cloud. Roam maps to ISO 42001 and SOC 2 controls via the Governance Evidence Pack — Roam does not certify and does not make a deployment compliant; your auditor judges fit for your scope. Book a 30-minute call from the Self-Hosted card above or email hello@roam-code.com.