Pricing

Everything on your own machine is free.

Unlimited usage on hardware you own — no token meters, no rate limits, no account. You pay only when other people get involved.

Per seat, billed monthly — cancel anytime.

Free

One person, one machine — and every GPU in it. No account, no meter, no expiry.

Full engine + entire model catalog
OpenAI + Anthropic-compatible API — point Cline, Continue, Aider or the OpenAI SDK at your own localhost:8080/v1
Bring your own model — load any Hugging Face repo
Developer & personal connectors — GitHub, Postgres, Gmail, Slack, arXiv & more, plus any custom MCP server you build
Local chat app + voice — nothing leaves your box
Built-in tools — web search, calculator, local memory & files

Download free

Enterprise

$20/seat/mo

For teams. Self-serve with a card — no sales call, no contract, cancel anytime.

Everything in Free, plus
Pool GPUs across machines — run 70B-class models no single card can hold (tensor parallel over your LAN)
Reach your box from anywhere — the hosted /v1 relay, outside your LAN
220+ business connectors — Salesforce, ServiceNow, Workday, Snowflake, Datadog & the full Security, Finance, HR & Healthcare catalog
Org, roles & seats — shared model cache, per-member access
Usage & activity dashboard
Email support

Enterprise Private

Custom

For regulated industries that must prove the data never left.

Everything in Enterprise, plus
True air-gap — offline mode + signed attestation
HIPAA (BAA available) & audit log export
SSO / SAML + SCIM + RBAC
On-prem control plane — zero dependency on baremetalrt.ai
Multi-site fleet management, SLA & named support

Talk to us →

No rate limits on any tier — you own the hardware, run it as hard as you like. Free is one machine. Pay when you pool GPUs across machines, or bring a team.

Compliance: GDPR · CCPA/CPRA · HIPAA (air-gap; BAA available) · PCI DSS (Stripe) · NIST AI RMF & EU AI Act-ready · SOC 2 & ISO 27001 in progress. Security & compliance →