Pricing
Start free. Scale when ready.
Two ways to start at $0 — no credit card required. Upgrade to dedicated instances when you need production reliability. No per-resolution fees, no surprise overages.
Free Sandbox
Try a live AI assistant right now — bring your own API key and chat instantly.
- Any model — Anthropic, OpenAI, Google, xAI, DeepSeek
- 50 messages per session · 30 min window
- Keys never stored — memory only
Community
Join our Discord, connect your Telegram bot, and get a live AI assistant — free.
- 1 shared instance · Telegram channel
- BYOK — use any model you want
- Community support on Discord
Launch production-ready assistants across core channels.
- 2 dedicated production instances
- Up to 3 channels per instance
- Any AI model (BYOK)
- Knowledge Base — train AI on your docs
- Embeddable chat widget
- Blog publishing — DEV.to + Hashnode
- 7-day log retention
Scale with multi-agent routing, browser actions and priority support.
- Everything in Starter, plus:
- 5 dedicated instances
- 5 channels per instance
- Up to 3 agents per instance (routing)
- Browser Actions — fill forms, navigate sites
- 30-day log retention
- Priority support SLA
For teams running production support agents at scale.
- Everything in Pro, plus:
- 10 dedicated instances
- 10 channels per instance
- 10 agents per instance
- 90-day log retention
- Dedicated support
- Custom usage reporting
All paid plans include AES-256-GCM token encryption, instance isolation, and cancel-anytime billing.
Two ways to pay for tokens
Platform Gateway or BYOK
Switch any time, per instance. The plan you pick decides capacity. The model mode decides who pays the LLM bill.
Platform LLM Gateway
We route every prompt to the cheapest model that fits the task — fast small models for simple replies, smarter models when you need depth. Pay metered credits at provider cost + a thin margin.
- No API keys to manage — top up credits, that's it
- Auto-routes across Anthropic, OpenAI, Google, xAI, DeepSeek
- Cheaper on average — small models handle ~70% of traffic
- Built-in fallbacks if a provider is down
- Per-conversation cost reporting
Powered by LLM Gateway — read why we partnered with them.
BYOK — Bring Your Own Key
Connect your own Anthropic, OpenAI, Google, xAI or DeepSeek API key. Provider bills you directly. Best when you have a contract rate, free credits, or a strict model requirement.
- Use your existing provider relationships and credits
- Pin one specific model (e.g. Claude Opus 4.6) every call
- Zero markup from us — you pay provider list price
- Keys encrypted with AES-256-GCM at rest
- Switch back to Gateway in one click
BYOK is available on every plan — including the free Sandbox.
Everything that's included
Every feature, illustrated.
One subscription unlocks the whole stack. The free Sandbox and Community plan use the same engine — they're just capacity-limited.
One-Click Deploy
Pick a model, connect a channel, hit deploy. Live in under 60 seconds — no Dockerfiles, no DevOps.
10+ Native Channels
WhatsApp, Telegram, Discord, Slack, iMessage, Signal, Microsoft Teams, Matrix, WebChat, Google Chat — all wired in.
Any AI Model
Claude, GPT, Gemini, Grok, DeepSeek, or your own. Switch in one click without redeploying.
Platform LLM Gateway
Auto-routes prompts to the cheapest model that fits the task. No keys to juggle, lower cost on average than a single pinned model.
Knowledge Base
Upload PDFs, CSVs, Markdown, sitemaps. Vector search retrieves context so answers stay grounded in your real docs.
Browser Actions
Give your bot a browser to fill forms, navigate dashboards, take screenshots, and complete real tasks.
Workflow Canvas
Drag triggers, conditions, AI actions and HTTP calls onto a canvas. Test runs show every step's input, output, and timing.
Embeddable Widget
Add an AI chat widget to any website with one script tag. Same agent powers your other channels.
Blog Publishing
Your AI writes and publishes articles to DEV.to and Hashnode from a single chat message.
Encrypted by Default
AES-256-GCM token encryption, JWT auth, per-instance rate limiting, full tenant isolation.
Always On
Managed infrastructure with health checks and auto-restarts. Status page with real incident history.
Multi-Agent Routing
Run several specialized agents inside one instance. Route by intent, channel, or keyword.
How Clawship pricing works
Platform Gateway or BYOK
Pick auto-routing through our LLM Gateway (cheapest model that fits the task) or bring your own provider key. Switch any time.
Pay only for tokens used
Gateway: metered against your credit balance at provider cost + thin margin. BYOK: provider bills you directly.
No per-resolution fees
Unlike Intercom Fin or Mintlify, we don't charge $0.40 – $0.99 per conversation. Volume doesn't punish you.
Compare plans
Every feature, side by side.
| Sandbox | Community | Starter | ProPopular | Business | |
|---|---|---|---|---|---|
| Capacity | |||||
| Dedicated instances | — | 1 shared | 2 | 5 | 10 |
| Channels per instance | — | Telegram | 3 | 5 | 10 |
| Agents per instance | 1 | 1 | 1 | 3 | 10 |
| Log retention | session | — | 7 days | 30 days | 90 days |
| Models & channels | |||||
| Platform LLM Gateway (auto-routing) | |||||
| BYOK — bring your own provider key | |||||
| Any model — Claude, GPT, Gemini, Grok, DeepSeek | |||||
| 10+ messaging channels | |||||
| Knowledge base (RAG) | |||||
| Embeddable chat widget | |||||
| Power features | |||||
| Browser actions | |||||
| Multi-agent routing | |||||
| Blog publishing (DEV.to + Hashnode) | |||||
| Workflow canvas | |||||
| Support | |||||
| Community Discord | |||||
| Email support | |||||
| Priority SLA | |||||
| Dedicated support | |||||
Compared to other AI agent platforms
Real numbers, not marketing math.
| Platform | Pricing model | Cost / 1k convs |
|---|---|---|
| Clawship | Flat $12 – $79/mo + BYOK tokens | ~$3 (BYOK) |
| Intercom Fin | $0.99 per "resolution" | ~$990 |
| Mintlify Chat | $0.40 per message | ~$400+ |
| Chatbase | $0.05 – $0.07 / message + seats | ~$50 – $70 |
Numbers reflect public pricing as of 2026. See vs Chatbase and vs Botpress for full breakdowns.
Pricing FAQ
Do you charge per resolution or per message like Intercom Fin?+
No. You pay a flat monthly subscription plus the cost of LLM tokens you actually use (BYOK). There are no per-resolution fees, no fake-resolution gaming, and no surprise $4,000 invoices.
What does BYOK mean?+
Bring Your Own Key. You connect your Anthropic, OpenAI, Google, xAI or DeepSeek API key. Clawship runs the assistant; the model bills you directly at provider rates. No 4× markup.
What's the difference between Platform mode and BYOK?+
Platform mode routes your prompts through our LLM Gateway, picking the cheapest model that fits each task and metering tokens against your credit balance. BYOK pins one specific model and the provider bills you directly. Platform is usually cheaper because ~70% of conversational traffic doesn't need a flagship model. BYOK wins when you have provider credits, a contract rate, or strict model requirements.
Can I cancel anytime?+
Yes. All paid plans are month-to-month and you can cancel from the dashboard at any time. No annual contracts, no exit fees.
What happens when I hit my plan limits?+
We notify you in-app before you hit a hard cap. You can upgrade or top up credits — no auto-overage charges without your consent.
Is there a free trial of paid plans?+
Both Free Sandbox and Community are zero-cost forever. They share the same engine as paid plans, so you can validate before upgrading. Paid plans don't currently include a separate trial.
Do you offer discounts for annual billing or non-profits?+
Yes — contact support for annual billing (≈ 2 months free) and non-profit / open-source discounts.
Where is my data stored?+
Tokens are AES-256-GCM encrypted at rest. Conversation logs are retained per your plan (7 / 30 / 90 days). You can export and delete data from the dashboard.
Can I self-host instead?+
Yes — Clawship is built on top of OpenClaw, which is open source. See /vs/managed-openclaw-hosting for the trade-off.
Need custom enterprise pricing? Contact us.