Every plan includes a dedicated cloud server. No hidden fees. No per-seat pricing. Deploy in ninety seconds.
For side projects and experimentation.
Get StartedThe best value for most users.
Get StartedFor production agents with API access.
Get StartedMulti-agent, SLA, and dedicated infra.
Need custom infrastructure?
Dedicated clusters, custom SLAs, white-label deployments, and volume pricing for teams deploying 10+ agents.
One user message sent to your agent. The agent's response — including any tool calls — counts as part of the same message.
Yes. Upgrade anytime from Dashboard → Settings. Your agent data (memories, sessions, config) is preserved during the server rebuild.
Buy message credits anytime — they're consumed automatically once your monthly quota is used up. Or per-message overage applies, from $0.01 (GPT-5 Nano) up to $0.05 (Claude Opus 5).
No. Choose "Use Lobstack AI" and our flagship Lobstack 1.2 model runs everything through the Lobstack Gateway — no keys, usage included in your plan. Prefer to bring your own provider key? BYOK is fully supported too.
Our flagship managed model. Behind one name, proprietary Token Intelligence routes every message in real time to the best engine for the job — fast and cheap for simple turns, flagship reasoning for hard ones — all through the OpenAI-compatible Lobstack Gateway.
Lobstack 1.2 plus all 30+ engines on Pro and above: Claude Opus 5, Claude Sonnet 5, GPT-5.6, Gemini 3.1 Pro, Grok 4.5, DeepSeek V4, Qwen 3.7, Mistral, and more. Starter includes Lobstack 1.2 and the standard (non-premium) models.
No. Lobstack handles all infrastructure — server provisioning, networking, security, monitoring. You describe what your agent should do, pick a model, and deploy.
No free tier, but you can start with Starter at $29/mo and upgrade anytime. All plans include a dedicated VM — there's real infrastructure behind every agent.