Simple, transparent pricing
Pay for seats.
Not tokens.
Run local LLMs and autonomous coding agents on your own hardware. Start free, upgrade when you need team-scale infrastructure. Your code never leaves your network unless you want it to.
Monthly cost
$0.00
Active seats
1
Local models
12
Data sent outbound
0 KB
Current plan
HobbyNo subscription · no telemetry · no API keys
Plans
Choose your tier.
Hobby is free forever. Pro unlocks power-user features. Team and Enterprise add collaboration, admin controls, and compliance.
Hobby
Run local LLMs privately on your own machine.
- Local-first, private inference
- 1 active device
- Core coding agent
- Community support
Pro
Power-user features for solo developers.
- Everything in Hobby
- Semantic code search
- Unlimited local models
- BYOK cloud models
- Priority email support
Team
Collaborate with shared endpoints, Git bridge, and RBAC.
- Everything in Pro
- Git bridge: branch, commit, push, PR
- Shared vLLM / SGLang endpoints
- Admin RBAC & seat lifecycle
- Usage metering & billing stats
- Priority support
Enterprise
Air-gapped appliance, SSO, compliance, and custom terms.
- Everything in Team
- Air-gapped / offline launch
- SSO & SCIM provisioning
- SOC 2 evidence export
- Data residency controls
- Custom contract & SLA
- Dedicated support
Compare
Feature breakdown.
Everything you need to pick the right plan. All tiers include local, private inference.
| Feature | Hobby | Pro | Team | Enterprise |
|---|---|---|---|---|
| Local inference | ||||
| Coding agent | ||||
| Active devices | 1 | Unlimited | Per seat | Unlimited |
| Semantic code search | ||||
| BYOK cloud models | ||||
| Git bridge | ||||
| Shared endpoints | ||||
| Admin RBAC | ||||
| Usage analytics | ||||
| Air-gapped deployment | ||||
| SSO & SCIM | ||||
| Custom SLA |
FAQ
Fair questions.
Asked by people who read source code before README badges.
Does the Hobby plan work fully offline?›
Yes. Once a model is downloaded, you can disconnect from the internet entirely. Chat, agent, and editor all keep working locally with no telemetry or account required.
What hardware do I need?›
An 8 GB VRAM GPU or better runs 7B-class models comfortably. The app auto-detects your hardware and picks a quantization that fits. Apple Silicon, NVIDIA CUDA, AMD ROCm, and Vulkan are all supported.
Can I switch between monthly and annual billing?›
Yes. You can change billing cadence at any time from your account settings. Annual plans are billed upfront and save roughly 20% compared to monthly.
What is the difference between Pro and Team?›
Pro is for individual power users. Team adds seat management, shared model endpoints, a Git bridge, admin RBAC, and usage metering for organizations.
Is Enterprise available air-gapped?›
Yes. Enterprise can be deployed entirely offline with SSO, SCIM, audit logs, and compliance exports. Contact sales for a custom contract, SLA, and data residency options.
Get started
Your hardware is more capable than you think.
Install in under a minute. Delete it just as fast — nothing lingers, nothing phones home.