Plans
Plans
BetterClaw has three self-serve plans — Free, Pro, and Business — plus a contact-sales Enterprise option. You can change plans at any time from the Plans & Billing page. Paying annually saves 20%.
PH3 at checkout for Pro at $16/mo for the first 3 months.Free
Best for trying BetterClaw with a single agent and your own API key. Closed to new signups while we add capacity — see the note above.
- 1 agent
- 500 credits / month
- 3 connectors (one account each)
- Bring your own API key (BYOK)
- 7-day memory retention
- Community support
Pro — $49/mo
For people shipping with agents every day. Billed monthly, or $39/mo billed annually.
Everything in Free, plus:
- 5 agents included (+$12/mo per extra agent)
- 12,000 credits / month
- Unlimited connectors, multi-account
- Ad integrations (Google, Meta, X)
- Residential proxy
- Managed LLM keys and BYOK
- 2 team seats (+$10/mo per extra seat)
- 90-day memory retention
- 24-hour email support
Business — $149/mo
For teams and agencies running many agents. Billed monthly, or $119/mo billed annually.
Everything in Pro, plus:
- 25 agents included (+$8/mo per extra agent)
- 40,000 credits / month
- Unlimited memory retention
- 5 team seats (+$8/mo per extra seat)
- Dedicated instance (2 vCPU / 2 GB)
- 12-hour priority support
Enterprise
For organisations that need SSO, custom retention, on-prem agents, or volume pricing. Enterprise is custom-quoted — reach out from the Plans page to talk to the team.

How credits work
Paid plans include a monthly pool of credits that meters two things:
- Agent runtime — 1 credit per minute your agent's machine is running.
- Managed LLM usage — when you use BetterClaw's managed model keys, the model spend is converted into credits. Usage on your own key (BYOK) is billed by your provider and doesn't draw from credits.
Credits reset each billing period. If you run low you can buy credit packs of 1,000 credits ($5 on Pro, $4 on Business), and the Plan & billing page shows where your credits are going.
Choosing a model
Regardless of plan, you choose which model powers each agent. On Free you supply your own API key (or sign in with your ChatGPT subscription); on paid plans you can use managed LLM keys or still bring your own. Anthropic leads the provider list, with Claude Fable 5 (native 1M context) at the top and Claude Sonnet 5 (also native 1M context) selected by default, followed by the Claude Opus, Sonnet, and Haiku families. See LLM configuration.