Start Free.
Scale When You're Ready.

Free is $0/month and includes 1 agent with 500 credits. Pro is $49/month for 5 agents and 12,000 credits/month. Business is $149/month for 25 agents and 40,000 credits/month. Billing annually drops Pro to $39/month ($468/year) and Business to $119/month ($1,428/year). Enterprise pricing is custom - contact sales.

Credits meter your agents’ work on the BetterClaw platform - task runs, connector calls, and proxy usage all draw from your monthly allowance. Free includes 500 credits/month, Pro includes 12,000, and Business includes 40,000. Credits reset every month. If you run out, extra credits are billed automatically in 1,000-credit packs ($5 on Pro, $4 on Business).

You can add agents beyond your plan’s limit for $12/agent/month on Pro or $8/agent/month on Business. Past about 12 extra agents on Pro, Business ($149/month for 25 agents) is cheaper - the dashboard flags this when you get close.

No. Your plan price covers the BetterClaw platform. Two things sit outside it: your AI model provider’s API usage, which you pay directly to them at their published rates with no markup from us, and any add-ons you deliberately choose (extra agents, extra credit packs, extra seats).

On the Free plan, yes - BYOK only. Pro and Business support both BYOK and managed keys, so you can let BetterClaw handle model access instead of wiring up your own provider account. BYOK works on every plan and carries zero markup.

It depends on your usage volume and which model you choose. As a rough guide, a typical agent handling moderate conversation volume on GPT-4o might cost $5–30/month in API usage. Lighter-weight models can bring that under $5/month. You control this entirely by choosing your model and managing usage through your provider’s dashboard.

Yes. Monthly plans have no contracts, no lock-in, and no cancellation fees. Cancel anytime from your dashboard and your agents keep running until the end of the current billing period. Annual plans are billed up front for the year.

You do not need one. The Free plan is free forever, not a countdown - 1 agent, 500 credits/month, no credit card. Upgrade when you hit the ceiling. Paid plans also carry a 7-day money-back guarantee.

OpenClaw is free and open source, but self-hosting carries real costs: a cloud server ($50–200/month), Docker setup (5–10 hours upfront), and ongoing DevOps maintenance (5–20 hours/week). Factoring engineering time at $75–150/hour, self-hosting typically runs $200–500+/month. BetterClaw Pro is $49/month with zero infrastructure overhead.

Yes. Annual billing saves 20% on both paid plans - Pro drops to $39/month ($468/year) and Business to $119/month ($1,428/year). For teams that need more than 25 agents, custom infrastructure, data residency, or a dedicated account manager, Enterprise pricing is negotiated directly. Contact sales.

Yes, on paid plans. Pro includes 2 team seats and Business includes 5, sharing one workspace with role-based access control (admin / member) so you control who can build, edit, and deploy agents. Extra seats are $10/seat/month on Pro and $8/seat/month on Business. The Free plan is single-user and has no team seats.

BetterClaw includes real-time health monitoring that continuously scans every agent’s status. If an anomaly is detected, the system automatically pauses the agent to prevent unintended behavior. You get full observability through your dashboard - logs, activity history, and diagnostics - so you can resolve issues quickly.

Yes, on every plan including Free. Every agent runs in a Docker-isolated sandbox with strict filesystem boundaries. All credentials are encrypted with AES-256. Agents can only access resources you explicitly allow through workspace scoping. All connections use cryptographic gateway pairing, and health monitoring auto-pauses agents if anomalies are detected.

No. BetterClaw is designed so anyone can deploy an agent without writing code or managing infrastructure. You create an agent, pick an AI model, connect your channels and tools, and it starts working. No config files, no Docker commands, no servers to provision. Most users deploy their first agent in under 60 seconds.