/ pricing
Start free. Scale when you need it.
Self-hosted Testhide is free forever — no limits on builds or agents. Cloud plans add managed infrastructure and higher AI quotas.
Free · Self-Hosted
Full Testhide on your own infrastructure. Your data stays on your server. No build limits, no agent limits, no expiry.
Install free →- Unlimited pipelines & agents
- 8 step types (pytest, Jest, .NET, Docker…)
- 1 concurrent build
- 7-day log retention
- AI Root-Cause Classifier
- AI Flakiness Predictor
- Win / Linux / macOS / Docker agents
- Community support
Cloud Starter
Billed monthly
Managed cloud. Zero infrastructure to maintain. Up and running in 5 minutes.
Contact sales →- 5 concurrent builds
- 30-day log retention
- 3 user seats
- 10,000 AI model runs / month
- Full 8-model AI dashboard
- Matrix builds
- Jira + Slack + Webhook alerts
- Email support (24h response)
Cloud Team
Billed monthly
Full platform for serious engineering teams. All 8 AI models, Investigator Agent, all integrations.
Contact sales →- 25 concurrent builds
- 1-year log retention
- 15 user seats
- 100,000 AI model runs / month
- All 8 AI models + Investigator Agent
- Matrix builds + TPS sharding
- Jira, Slack, Teams, Webhooks
- vCenter / ESXi VM agents
- Priority support · 4h SLA
Enterprise
Annual contract · custom SLA
On-prem or air-gapped. SSO/SAML. White-label. Custom AI training. Dedicated success manager.
Talk to us →- Unlimited concurrent builds
- Unlimited log retention
- Unlimited users & agents
- Unlimited AI model runs
- On-prem / air-gapped deployment
- SSO / SAML / LDAP / Active Directory
- White-labeling
- Custom AI model training on your data
- Dedicated success manager
- Custom SLA + SOC 2 / GDPR support
* Self-hosted Free: unlimited builds, agents, and AI model runs — they run on your own hardware, so there's no cloud quota — with a single-threaded scheduler. Cloud tiers meter AI runs (Starter 10,000 / Team 100,000 per month) and add parallel execution + managed infrastructure.
/ compare
Full feature comparison.
| Feature | Free | Starter | Team | Enterprise |
|---|---|---|---|---|
| Build Platform | ||||
| Concurrent builds | 1 | 5 | 25 | Unlimited |
| Log retention | 7 days | 30 days | 1 year | Unlimited |
| User seats | 1 | 3 | 15 | Unlimited |
| 8 step types (pytest, .NET, Docker…) | ✓ | ✓ | ✓ | ✓ |
| Matrix builds | — | ✓ | ✓ | ✓ |
| TPS parallel sharding | — | — | ✓ | ✓ |
| vCenter / ESXi VM agents | — | — | ✓ | ✓ |
| Jenkins pipeline import | ✓ | ✓ | ✓ | ✓ |
| Build artifacts | ✓ | ✓ | ✓ | ✓ |
| AI Intelligence | ||||
| AI model runs / month | Unlimited* | 10,000 | 100,000 | Unlimited |
| Root-Cause Classifier (DistilBERT) | ✓ | ✓ | ✓ | ✓ |
| Flakiness Predictor (41 features) | ✓ | ✓ | ✓ | ✓ |
| Full 8-model AI dashboard | — | ✓ | ✓ | ✓ |
| Failure Retriever (FAISS) | — | ✓ | ✓ | ✓ |
| Visual Diff Analyzer (YOLOv8) | — | ✓ | ✓ | ✓ |
| AI Investigator Agent (LLM) | — | — | ✓ | ✓ |
| Custom AI model training | — | — | — | ✓ |
| Integrations | ||||
| REST API v3 (246+ endpoints) | ✓ | ✓ | ✓ | ✓ |
| Jira auto-linking | — | ✓ | ✓ | ✓ |
| Slack notifications | — | ✓ | ✓ | ✓ |
| Microsoft Teams | — | — | ✓ | ✓ |
| Custom webhooks | — | ✓ | ✓ | ✓ |
| Security & Compliance | ||||
| Data stays on your infra | Always | Cloud | Cloud | On-prem |
| RBAC + PAT tokens (15 scopes) | ✓ | ✓ | ✓ | ✓ |
| SSO / SAML / LDAP | — | — | — | ✓ |
| Air-gapped / offline AI | ✓ | — | — | ✓ |
| White-labeling | — | — | — | ✓ |
| SOC 2 / GDPR compliance support | — | — | — | ✓ |
| Support | ||||
| Community (GitHub / Docs) | ✓ | ✓ | ✓ | ✓ |
| Email support | — | 24h | 4h SLA | Custom |
| Dedicated success manager | — | — | — | ✓ |
/ faq
Common questions.
Is the self-hosted Free tier really free forever?
Yes. Self-hosted Testhide has no expiry, no feature flags, and no call-home requirement for core functionality. You can run unlimited builds and connect unlimited agents on your own hardware with no payment ever required. The only difference from Cloud is that you manage your own infrastructure and get 1 concurrent build slot on the scheduler.
What counts as an "AI model run"?
One AI model run = one inference call to any of the 8 built-in models (root-cause classifier, flakiness predictor, FAISS retriever, OOD detector, visual diff, log miner, bug linker, emerging issues). Investigator Agent (LLM) calls to external providers like Gemini or GPT-4 are billed separately through your own API key — Testhide never marks up those costs.
Can I exceed the AI model run quota?
On Cloud plans, when you hit the monthly quota, AI model inference pauses for the rest of that billing cycle — builds continue running normally, they just won't get AI triage labels until the next cycle. We notify you at 80% and 100% of quota. You can always upgrade mid-month for a prorated charge.
Monthly vs yearly billing?
Monthly billing renews automatically each month. Yearly is billed upfront with a 15% discount — Starter drops from $49 to $42/mo ($504/yr), Team drops from $299 to $254/mo ($3,048/yr). You can switch between monthly and yearly at any renewal date.
Can I upgrade or downgrade mid-month?
Yes. Upgrades take effect immediately and the remaining days are prorated. Downgrades take effect at the end of the current billing cycle.
What's the refund policy?
30-day money-back guarantee for new Cloud subscriptions. After 30 days, refunds are evaluated case by case. Contact us at the address on the About page.
Is Cloud available now?
Cloud plans are in early access — invite-only while we scale the managed infrastructure. Contact us to get on the waitlist. Self-hosted Free is available right now with a 5-minute Docker install.
How is Enterprise different from Cloud Team?
Enterprise adds: on-premises or air-gapped deployment (your data never leaves your network), SSO/SAML/LDAP, white-labeling, custom AI model training on your historical failure data, a dedicated success manager, and a custom SLA. It's also the path for regulated industries (finance, healthcare, defence) that can't use managed cloud.
/ get started
Self-hosted. Free forever. Up in 5 minutes.
No cloud account. No credit card. Docker Compose installs the full platform — builds, AI models, dashboard — in one command.