Start with one feature. Add capacity as you use RedCrown across more work. Free covers one active AI feature and one hosted proof. Paid plans start at Starter, $119/mo billed monthly, or $1,188 billed once a year (the equivalent of $99/mo), and scale with the number of active AI features.
An active AI feature is an experiment in your workspace that is neither archived nor deleted. Re-running an existing feature does not use another slot. Archiving frees the slot and keeps the history. Free's separate lifetime cap of one hosted proof is never returned by archiving.
One feature, its cases, its references, and the requirements it is checked against. A ticket router is one feature. A clinical transcription step is another. Every run of that feature, a new candidate, a new model, a re-run after a provider update, appends to the same record. redcrown push --experiment <id> appends a run to an existing feature and uses no new slot.
Creating an experiment in the app, the CLI or the API. Scoring supplied outputs (Advanced, Upload, Score supplied outputs) to a new feature creates one. Scoring to an existing feature adds a version and creates no new feature. An offline redcrown score or redcrown check creates nothing, because nothing is uploaded.
Archiving takes a feature out of your active count immediately, on every plan, Free included. Its history, receipts, and share links stay readable. Restore it later and it counts against your plan again. Free's separate lifetime cap of one hosted proof is a different thing: archiving never returns it. Evaluation runs, model comparisons, dataset rows, report viewers, reviewers, and rerun history are uncapped on every plan.
| Free | Starter | Growth | Scale | Enterprise | |
|---|---|---|---|---|---|
| Active AI features | 1 | 3 | 10 | 30 | Custom |
| Evaluation runs, bring your own keys | Unlimited | Unlimited | Unlimited | Unlimited | Unlimited |
| Hosted, shareable proof links | 1, for the life of the account | Unlimited | Unlimited | Unlimited | Unlimited |
| Blinded reviewers | Unlimited | Unlimited | Unlimited | Unlimited | Unlimited |
| Re-run on demand + version history | ✓ | ✓ | ✓ | ✓ | ✓ |
| Workspaces, invite collaborators | Unlimited | Unlimited | Unlimited | Unlimited | Unlimited |
| Priority support | – | – | ✓ | ✓ | ✓ |
| Self-hosted / VPC | – | – | – | – | ✓ |
| Signed attestation on any proof | ✓ | ✓ | ✓ | ✓ | ✓ |
An active AI feature is an experiment in your workspace that is neither archived nor deleted. Re-running an existing feature does not use another slot. Scoring supplied outputs (Advanced, Upload, Score supplied outputs) to a new feature creates one. Scoring to an existing feature adds a version and uses no new slot. redcrown push --experiment <id> appends a run to an existing feature. Archiving frees the slot and keeps the history. Free's separate lifetime cap of one hosted proof is never returned by archiving.
No. You bring your own provider keys and the calls run under your accounts at your rates. RedCrown takes no margin on those calls and no revenue share from any model provider. Third-party inference charges are yours and are separate from any RedCrown plan. That is also how the neutrality is structured.
Yes. Free includes 1 active AI feature and 1 hosted proof for the life of the account, with no time limit and no credit card. Evaluation runs and model comparisons are unlimited on Free, with your own keys. The first hosted proof link your account creates also starts the activation promotion: 30 days of Growth entitlements, granted once per organization, with no card. When the 30 days end, nothing is charged; if the workspace holds more than Free's 1 active AI feature, the least recently used features are archived down to the cap, so the one you worked on last stays active, and archived features stay readable.
No. redcrown score and redcrown check run offline with no account at all. The app is free for one active AI feature with no card and no sales call, and you upload only the results you choose.
Nothing breaks and nothing is deleted. Archive an existing feature to free a slot, or upgrade to a plan with more active AI features. Archived features stay readable, including their share links; restore one and it counts against your plan again. That is separate from Free's lifetime limit of one hosted proof, which archiving does not affect.
How many active AI features you can keep at once: 3 on Starter, 10 on Growth, 30 on Scale. Growth and Scale add priority support and the proxy capture path (API and MCP setup). Unlimited runs, hosted proofs, and reviewers are the same on every paid plan.
It depends on the path. redcrown score and redcrown check send nothing anywhere. A native evaluation, local or hosted, sends requests to the model providers you configure, on your keys. Uploading a run and creating a share link are two separate, deliberate actions, and a public share link shows inputs, outputs and references. Data is siloed per workspace, and share links revoke when you say so. Your provider keys, uploaded eval inputs, and proof-link tokens are encrypted at rest under the application key-encryption key that also protects provider keys; eval inputs use a separate PHI key when one is configured. Stored run reports, which hold the model outputs and the inputs they were run on, sit in the database without that additional layer, so treat a hosted run as you would any other cloud-stored copy of the data. RedCrown holds no formal certification (no SOC 2, no ISO 27001) and does not claim any. Details on the Security & Trust page.
No. RedCrown holds no SOC 2 report and no ISO 27001 certificate, and no audit is under way. Our platform providers hold their own: Railway states SOC 2 Type II, SOC 3 and HIPAA, Supabase states SOC 2 Type 2, ISO 27001 and HIPAA, and Netlify lists SOC 2, ISO 27001 and PCI DSS. Those attestations cover their platforms, not RedCrown, and we will not present them as ours. If your procurement needs a report from us, tell us and we will scope it with you. The Security & Trust page lists every subprocessor and what it publishes.
RedCrown takes no payment, revenue share, or placement fee from any model provider, and takes no margin on the calls you run. You bring your own provider keys, so the calls run under your accounts at your rates. Scoring runs on a pinned, versioned methodology, and every report records the version that produced it. Read the full neutrality structure.
It would be, on its own. So score against your own labeled references, or send the run to a blinded human reviewer who votes A against B without seeing model identities. When you do use a judge, it is pinned and versioned, and every verdict links to the receipt it was based on so you can check the grading yourself.
Score the candidates against your current model's output instead of a labeled reference. That shows which candidates match what you ship today and what each would cost at your volume. It does not establish that the current output is correct. With labels, RedCrown scores against those and reports quality as well.
For hands-on onboarding, security review, or procurement. We reply by email.
Free, on your own feature, in your browser or your terminal. No sales call, no credit card.