Open Core stays capable
Connected desktop work, local models, self-hosted deployments and supported provider/plugin paths are not sold back to you as a daily message allowance.
Use HAL without a membership message counter when the work runs on your own connected computer, local models, self-hosted providers or supported tools. Memberships fund and prioritize HAL-managed compute, production capacity, media infrastructure and direct technical support.
HAL is being built as an open, connected system. Your hardware and providers should remain useful without a subscription. Paid plans exist to add managed infrastructure and service capacity when you want HAL to carry more of the compute burden.
Connected desktop work, local models, self-hosted deployments and supported provider/plugin paths are not sold back to you as a daily message allowance.
Free access can use community hosted capacity when available. Paid plans receive progressively larger hosted burst capacity and service priority during congestion.
Revenue supports GPU/CPU upgrades, provider bills, servers, storage, background execution, voice/media compute, reliability, development and support.
Every paid tier includes Open Core. The difference is how much HAL-managed infrastructure, queue priority, media capacity and direct support is attached to your account—not whether you are allowed to keep talking to HAL.
“No membership cap” refers to work performed on compute you supply or control. HAL-managed hosted services still need fair-use, abuse-prevention, provider and capacity controls. Paid plans increase managed capacity and priority; they do not promise infinite third-party GPU or model spend.
1 credit = $0.001 of verified HAL-managed provider/GPU cost. Credits are internal service accounting units—not cash, crypto, property or transferable value. Local, connected, self-hosted and user-provider compute does not debit this pool. Unknown provider cost is never guessed into a charge.
HAL’s value is the connected system around the models: coding, research, computer connection, tools, media, voice, automations and a shared website/Desktop workspace.
Fast, Build, Expert and Council modes; connected computer workflows; local models; supported plugins/MCP tools; files and project context; website/Desktop continuity.
Music generation, queue priority, longer output retention on paid plans, premium-quality voice paths when healthy, and the broader Studio/media pipeline as capacity comes online.
Membership state, PayPal billing, account/email linkage, priority support/review and increasingly managed infrastructure for people who want HAL to carry more of the operational burden.
These are active development targets funded by membership revenue and will only become guaranteed plan entitlements when the corresponding metering and capacity controls are production-ready.
Roadmap items are not guaranteed by a current plan until they appear in that plan’s “included” list. This keeps the offer honest while making the direction explicit.
Revenue helps cover model/provider bills, GPU/CPU and workstation upgrades, cloud servers, storage and bandwidth, reliability work, continued development, promotion/outreach, and operating margin / creator compensation.
Plans do not provide equity, ownership or profit-sharing. Hosted capacity depends on provider availability, system health, fair-use and safety controls. Scheduled review time on Studio, Infrastructure and Foundry is subject to mutual scheduling, is advisory/review time rather than a guaranteed software deliverable, and does not roll over unless separately agreed. Recurring billing is processed by PayPal.