One flat price.
Zero meters running.
Cloud AI charges you more the more you build. Veylaro runs on your hardware, so heavy usage costs us nothing — and we price it that way.
Free
- Laro Lite — full local weights
- First 30 days unlimited at launch
- 50 agent messages per week — resets every Monday
- Chat, file & image understanding
- Single-project memory
- Community support
- Unlimited usage
- Local API access
- Commercial license
Pro
- Unlimited usage — run it 24/7
- Full agentic coding suite
- Advanced long-term memory
- Reasoning visibility & transcripts
- Local API endpoint
- Commercial use license
- Priority model updates
- Priority support
Team
- Everything in Pro, per seat ×5
- Shared team memory & conventions
- Centralized license management
- Admin dashboard & usage insights
- Policy controls per workspace
- Priority support with SLA
Enterprise
- Self-hosted fleet deployment
- SSO / SAML & SCIM provisioning
- Air-gapped environments
- Custom model fine-tunes
- Security review & DPA
- Dedicated support engineer
Model inference runs locally. Prices shown are USD; checkout also offers NZD pricing. Pro is US$29/month or US$290/year (about US$24.17/month billed annually). The Free plan needs only an account; your card is never asked for. Launch-month unlimited applies before the normal plan limits begin.
The math cloud vendors
hope you never do.
A heavy agentic workflow on cloud AI can burn through hundreds of dollars of metered usage a month — and the bill grows with your ambition. Veylaro is a flat $29. Use it 10× more and it costs exactly the same.
Illustrative comparison of typical metered cloud-agent spend vs. Veylaro Pro flat pricing.
Fair questions.
Straight answers.
Free includes local Laro Lite so you can evaluate Veylaro on your own projects. Paid plans change allowances and features; model tiers have different hardware requirements and must be qualified separately. Check the download and benchmark pages before choosing.
The subscription funds the agent software, model qualification, updates and support. Local inference uses your hardware. The current preview uses upstream Gemma weights; a separately qualified Veylaro fine-tune is not claimed.
Laro Lite needs 8 GB RAM. Laro Med needs 12 GB and is recommended at 16 GB. Laro Max needs 24 GB and is recommended at 32 GB. Veylaro selects conservatively so the app stays responsive.
Local inference, project indexing, memory and transcripts stay on-device. Accounts, billing, downloads, updates and optional live search use the network. Search queries or explicitly selected online actions may transmit their inputs; do not include secrets. Offline use does not provide live search or online account services. See the privacy policy for the boundary.
There is no meter because there is nothing to meter. Compute is yours, so we don't count your tokens — a 12-hour overnight refactor costs us the same as an idle app: nothing. This is the structural advantage of local AI, and we're passing it on.
Pro, Team and Enterprise include a full commercial license. The Free tier is for personal projects and evaluation.
At launch, every new account gets 30 days of unlimited use across the plan it selects. No card is required for Free. Paid purchases also have a 30-day refund window after the first payment.
Start free. Stay unlimited.
Download the app, meet the model, and upgrade when it earns it.