Pricing
Pay $99.
Hire a workforce.
Or pay $10K to have us build it for you. Either way, you stop pretending VAs and consultants are scaling your operation.
Self-serve
Personal Dashboard + employees
Plus per-employee monthly pricing. Most operators run 3-7 employees at $79-$2,997/month each, depending on the role.
+ $99/mo per additional brand. Hired employees work across every brand for the same price.
- Onboarding intelligence — interviews you, writes your brand + ops knowledge files
- Weekly intelligence updates — every employee gets smarter automatically
- Powered by Claude Opus 4.7 + Sonnet 4.6 + Haiku 4.5 (cost-optimized routing)
- Personal Dashboard with telemetry, content pipeline, calendar
- Brand-leak validator (deterministic regex) — runs on every output
- Groundedness + instruction-follow LLM judges on demand
- Eval harness with golden cases + Sunday regression alerts
- Daily Brief at 7 AM + drift detection at 6 AM
- Off-host encrypted backup
- Multi-brand support — assign one employee to multiple brands without paying twice
- Multi-channel alert delivery (Telegram + webhook + email)
- Provenance tracking — every action linked to its inputs and outputs
No setup fees. Cancel anytime.
White-glove
Custom dashboards
For executives running multi-million-dollar businesses with concrete operating problems and a need for hands-on partnership.
- Dedicated infrastructure — separate VPS, separate Clerk org, separate database
- Custom agent training on your historical data, brand voice, and proprietary workflows
- White-glove onboarding led by a senior operator (not an automated intake)
- Integration build-outs against your in-house systems (legacy databases, custom CRMs, internal APIs)
- Quarterly strategy reviews — your agent stack reviewed against business outcomes
- Direct line to engineering for high-priority changes and integrations
- SOC 2 Type 2 path support, security questionnaire response, DPA negotiation
- Per-tenant audit logs, encryption, and customer-controlled data residency
What you get for $99
The base layer that makes the agents work.
Most marketplaces sell agents. We sell the base layer that turns generic agents into your agents — and price it to be obvious.
Onboarding intelligence
When you sign up, an intake agent interviews you about your business(es), products, voice, customers, and goals. It ingests any documents, websites, or recordings you provide and produces the brand and operations knowledge files that every other agent reads at runtime. This is the layer that turns a generic agent into your agent.
Weekly intelligence updates
Every agent in the catalog ships improvements weekly. Prompt revisions, new tool integrations, security patches, and quality refinements flow automatically. You always run the best-of-class version. Regression tests block any update that would degrade quality on your golden cases.
Claude Opus 4.7 + Sonnet 4.6 + Haiku 4.5
Each agent routes to the right Claude model for the job. Opus 4.7 handles strategic synthesis and high-stakes reasoning. Sonnet 4.6 handles most production work. Haiku 4.5 handles high-volume classification and triage. The cost difference is up to 75× — routing intelligently is how the marketplace stays affordable.
Mission Control dashboard
The web app. Content pipeline (review and approve agent output), calendar, prospect board, brand pages, system logs, and the agent marketplace itself. Mobile-friendly. Auth-protected.
Brand-leak validator (regex)
Every piece of content an agent produces gets checked against a per-brand keyword guard. Catches cross-brand contamination (e.g., a competitor's product mentioned in your copy), banned phrases, FTC-risky language, and tech-stack name-drops. Runs in milliseconds. Free with the base subscription — no LLM cost.
Groundedness + instruction-follow LLM judges
Two on-demand LLM checks for any content piece. Groundedness verifies that factual claims are supported by your brand master file. Instruction-follow verifies the content follows your stated voice rules and required disclaimers. ~$0.001 per check via Haiku 4.5.
Eval harness + Sunday regression alerts
Every Sunday at 5 AM, the entire golden test set runs against the current agent prompts. If quality drops vs. the prior week, you get an alert. This is what prevents prompt drift and lets you trust agent updates without manual review.
Daily Brief + drift detection
Daily 7 AM Telegram message: what your agents did in the last 24 hours, what's pending review, anything flagged. Daily 6 AM drift check looks at the prior week vs. the week before — if any metric is moving the wrong direction, you find out before it becomes a problem.
Off-host encrypted backup to Backblaze B2
Daily encrypted snapshot of your data, brand files, agent memory, and configuration to Backblaze B2. Restore drill validated. If your VPS catches fire, you're back online in under an hour with everything intact.
Per-brand content separation
If you operate multiple brands or businesses, the agents auto-detect which brand a task belongs to and load the right knowledge files. No accidental cross-pollination between businesses with different audiences or voices.
Multi-channel alert delivery
Alerts route to wherever you actually live — Telegram, Slack/webhook, or email. Critical alerts bypass quiet hours. Routine alerts get batched into the morning brief.
Provenance tracking
Every piece of content shows which agent produced it, what inputs it had, when each judge ran, and how it was edited or approved. When something looks off, you can trace it back to the exact agent run that produced it.
Frequently asked
Pricing questions, answered.
- Why a $99 base subscription?
- The base subscription is what makes every agent yours instead of generic. It includes the onboarding intelligence (interviews you, ingests your materials, writes your brand and ops knowledge files), the brand-leak validator (runs on every output), the eval harness (Sunday regression testing), Claude Opus 4.7 reasoning, and the dashboard surface. Without that layer, agents would be configurable assistants — the layer is what turns them into a workforce.
- How much does a typical customer spend?
- Most operators run 3–7 agents on top of the base. Average self-serve ARPU is $400–800/month all-in. Heavier operators with custom dashboards start at $10K per engagement.
- What's the difference between self-serve and white-glove?
- Self-serve customers configure their stack from the marketplace, run on shared infrastructure, and onboard via the automated onboarding intelligence. White-glove customers get dedicated infrastructure (separate VPS, separate Clerk org), a senior operator running their onboarding by hand, custom integrations against their in-house systems, and quarterly strategy reviews. White-glove starts at $10K per engagement; self-serve starts at $99/month.
- How does white-glove compare to hiring an AI consultant?
- AI consultants run $300–800/hour and produce custom workflows that need maintenance. We deliver pre-built agents that ship weekly improvements automatically. The white-glove tier gives you the same hands-on attention with a workforce that compounds — the consultant scope ends; the workforce keeps shipping.
- What happens if quality drops on an agent update?
- Every agent has a golden test set. Updates run against your golden cases before deployment. If quality regresses, you get an alert and the update is held back. You can also pin to a specific agent version if you want manual control over upgrades.
- Can I cancel anytime?
- Yes. Self-serve subscriptions cancel from the dashboard with one click. White-glove engagements have agreed-upon terms but no lock-in beyond the engagement period. Your data is yours and exports cleanly on cancellation.