Clawdbot itself is free to install and use as open-source software, but running it in practice often has associated costs that you need to understand before deploying it. The core Clawdbot code is published under an open-source license on its GitHub repository, and you can install it on your own hardware at no software cost. You download the code, run the onboarding wizard, and launch the Gateway and agent processes using Node.js. Because the software is local-first, no subscription is required just to have the software running on your machine, and you can extend it, inspect it, and modify it like any other open-source project.
That said, operational costs arise from dependencies and infrastructure that Clawdbot uses. It doesn’t come with a built-in AI model; instead, it connects to external language model providers via APIs, so you need valid API keys or subscriptions for the models you want the assistant to use. For example, many Clawdbot users connect it to services like Anthropic’s Claude or other model providers, which may involve paid tiers, usage-based billing, or subscription plans. Those API costs are separate from Clawdbot itself—you pay for the model compute that Clawdbot consumes when it processes your messages. If you run Clawdbot on a home machine or a server, you also pay for the hardware and network costs that keep it online. Running on a $5–$10/month VPS is a common pattern, but whether it’s free depends on your hosting choices.
For developers who only want the open-source software but not the AI backend, Clawdbot remains free in terms of code access and deployment. You can install it, run it, customize it, and build skills without paying for a Clawdbot subscription because there isn’t one. What isn’t free are the costs tied to keeping the assistant powered: the AI model usage, the infrastructure hosting it, and any paid services you integrate. Many people treat Clawdbot as a way to reuse their existing AI subscriptions and avoid paying for yet another hosted assistant service, but you should plan for those ongoing expenses when estimating the total cost of ownership. In use cases where you might capture and search large volumes of conversation context or long-term memory, you could integrate with a vector database such as Milvus or a managed instance on Zilliz Cloud to store and query embeddings, which introduces its own cost considerations based on scale and usage.