Loading
Loading
Connect to our cloud for self-serve AI tools, or deploy a sovereign platform with custom agents built for your business.
Run your own AI agent locally or deploy agent-powered apps with awkit. Every tier runs with your own models — Ollama, vLLM, or cloud APIs.
CLI access only
0 agents, 0 tools
4 packs included
1 agent, 120 tools, 10 skills
14 packs included
1 agent, 450 tools, 50 skills, 2 workspaces
18 packs included
5 agents, 650 tools, 80 skills, 5 workspaces
All 42 packs
Unlimited
All 42 packs
Unlimited
Add agent packs (Iris, Hydra, Athena, Hera, Saga) to any paid tier starting at $19/mo each. Browse the full catalog to mix and match.
Browse Pack StoreStandalone products you can add to any plan. Each runs on your infrastructure with zero-knowledge encryption.
| Bundle | Bundle Price | Separate Price | Savings |
|---|---|---|---|
| SecurityStack | $79/mo | $97/mo | 19% |
| SecOps Suite | $199/mo | $363/mo | 45% |
| Platform Plans | Sovereign Platform | |
|---|---|---|
| Inference | Token-metered | Unlimited 24/7 |
| Data | Tenant-isolated cloud | Your hardware, full sovereignty |
| Agents | Use ours via API | Custom-built for your workflows |
| Support | Community + docs | Direct line, no ticket queue |
| Setup | Self-serve, instant | 3-month deployment with us |
| Best for | Developers, indie hackers, builders | Businesses, teams, firms |
AitherOS is open source. Deploy the full stack on your own hardware with Docker Compose — no subscription required.
Already running Ollama, vLLM, or paying for OpenAI? Skip our inference — pay only for our MCP tools. Catalog & per-tool credit pricing.
Tokens are the unified currency for BYOC tool-credit plans. They cover LLM inference, image generation, agent execution, and GPU time. Purchased tokens never expire.
No. Sovereign deployments include unlimited inference — no metering, no per-request costs. You run agents 24/7 without counting tokens.
Platform plans are self-serve: pick a tier, run your own agent locally or deploy with awkit. Sovereign is a managed deployment: we build custom agents for your business, deploy on dedicated infrastructure, and maintain it. Platform plans are for developers. Sovereign is for teams and firms.
Yes. Many customers start on Cloud to evaluate the tools, then move to Sovereign when they need custom agents, unlimited inference, or data sovereignty.
The first three sovereign customers get pricing locked at roughly 50-60% below standard rates for 24 months. In exchange, you commit to quarterly feedback sessions and reference rights. Two slots are currently open.
Yes. BYOC tool credits accept Solana (SOL). Sovereign invoicing is standard net-30.
The Free tier is perpetual — no credit card required. For Sovereign, the initial deployment is scoped and priced upfront; if it doesn't earn its keep by month three, you walk.
Yes. Most products offer a self-hosted option with Docker. You get a one-time license fee, run it on your own hardware, and your data never leaves your network. See individual product pages for Docker images and system requirements.
Only for AI inference products (Canvas, Voice & Vision, Training Pipeline). Security, business, and communication products run on any CPU. The full AitherOS stack runs on a single RTX 5090.
Always. Every product uses zero-knowledge encryption. We cannot see your data even in managed deployments. Air-gapped deployment is available for Sovereign Node customers.
SecurityStack (CA + Directory + Identity, save 19-33%) and SecOps Suite (9 security services, save 45%+). Comms Hub and Media Studio are also bundled offerings. See the Product Add-ons section above.
Run the whole platform on your own hardware. Offline licence, no phone-home, air-gap capable.
$299/mo
Minimal sovereign node: core agent runtime + node/mesh enrollment, CPU-only. ~62 services. Self-service ISO/VHDX you boot and it joins your tenant overlay.
$899/mo
Standard sovereign node: agent runtime + memory/knowledge stack + mesh, CPU-only. ~102 services.
$1,500/mo
GPU sovereign node: full agent + knowledge stack + local vLLM inference (VLLM/Swap/Embeddings). ~108 services. Requires a GPU host.
$3,500/mo
GPU sovereign node with the reasoning model stack (qwen3.6-class) + cognition services. ~139 services. Requires a capable GPU host.
$4,167/mo
Build any-tier appliances yourself on your own bootc host under an annual build license. For air-gapped / fully-sovereign deployments; no managed CI, no data leaves your network.
$12,000/mo
The complete sovereign fleet: every service group, GPU inference, cognition, creative, and ops. ~272 services. Multi-GPU host recommended.
Support is priced as a percentage of licence with the floors below. Deployment and hardware are quoted against a written scope.
$1,500/mo
Next-business-day response, business hours, email + portal. Includes version upgrades, security advisories and known-issue workarounds. Priced at 15% of annual licence with this monthly floor; accounts above ~$120k/yr of licence are quoted on the percentage.
$2,400
One engineering day, remote, senior specialist. On-site adds $1,500/day plus travel at cost + 10%.
from $3,500
Mini-PC, CPU-only, 64GB. Assembled, burned in, imaged with your appliance, asset-tagged and shipped. 12-month return-to-base warranty. From — final quote follows current BOM. Licence sold separately.
$5,000/mo
4h P1 response, 24x7 for P1, named support contact, quarterly review and upgrade assistance. Priced at 22% of annual licence with this monthly floor; accounts above ~$273k/yr of licence are quoted on the percentage.
$7,500
One node, installed remotely with you, handover session and two weeks of follow-up. About a week.
$12,000/mo
Five engineering days a month with reserved capacity. Unused days roll over one month only — the capacity was already held for you.
from $12,000
Single GPU (5090-class), 128GB. Assembled, burned in, imaged, asset-tagged and shipped. From — final quote follows current BOM. Licence sold separately.
$15,000/mo
1h P1 response 24x7, a named engineer who knows your deployment, quarterly architecture review, upgrade rehearsal and an RCA within 5 business days for every P1. Priced at 30% of annual licence with this monthly floor; accounts above ~$600k/yr are quoted on the percentage.
$22,000
Ten engineering days against a written scope, at a small discount to the day rate for committed work. Change requests bill at the day rate.
$25,000
Up to five nodes, mesh enrolment, identity integration, a written runbook and one training day. Two to three weeks.
from $38,000
Dual GPU, rack-mount, redundant PSU, IPMI. Assembled, burned in, imaged and shipped. From — final quote follows current BOM. Licence separate.
$75,000
HA across sites, DR rehearsal, SSO/SAML integration, custom runbooks and admin training. Six to eight weeks.
from $120,000
Fully offline install: no phone-home, offline licence issuance, a compliance evidence pack and customer-witnessed on-site delivery. Priced from — scope and travel are quoted.
from $150,000
Four nodes with switching and PDU, built, shipped and installed on site. From — scope, BOM and travel are quoted.