Questions, answered plainly

Eddy is an AI platform your organization owns: one shared memory across your meetings, email, and systems that helps your team prepare, answers from your company's context, and follows through on the work that comes out of each conversation.

The Eddy software is open source, the models we deploy by default are open-weight files you hold, and your data and configuration stay under your control, whether Eddy runs on a dedicated server we operate for you or on hardware you own. Because you hold the code, models, and data, a change in our pricing or terms can't take the system away from you. Ownership is about control, not about where the machine sits on day one.

Your data, Eddy's memory of your organization, your configuration, and the documentation for your deployment. You can request a full export at any time. The details are set out in your agreement.

Owning the system doesn't mean it runs in isolation. Eddy needs somewhere to run: a dedicated server we operate, cloud hosting, or hardware you own. It connects to the tools you choose, such as email, calendar, meetings, phone, and documents, so those providers' terms still apply. If you route some work to a frontier model provider, that work runs under their terms. We document every dependency in your deployment so nothing is hidden.

Not by us. Open-weight models running on infrastructure you control don't send your data back to the company that made them. If you choose to route some work to a frontier provider, that provider's terms apply; where they offer zero-data-retention agreements, we'll configure them. If you want to keep outside parties to a minimum, models can run on hardware you control, though your hosting, network, and connected-tool providers are still involved.

Eddy acts with the permissions of the person asking. If you can't open a file, Eddy won't show it to you or use it in your answer.

You decide where the line sits. Actions such as sending a message outside the company, changing a record, or updating a connected system can be set to require a person's approval first. Answers cite their sources, so your people can click through and check.

During Ongoing Managed Care, we do: updates to models and features, monitoring, and support when you need it. If you'd rather run it yourself or with an IT partner you trust, you can. In that case, maintenance becomes your responsibility or theirs, and we hand over the documentation needed to do it.

The system remains yours and can keep running. Updates, monitoring, and support from us stop. Costs that continue include infrastructure (hosting, or your own hardware and power), any usage of outside model providers, and the fees for the tools Eddy connects to. Keeping it secure and current becomes your team's or your IT partner's job.

You keep the code, models, data, and documentation, so another qualified team can operate and maintain your deployment. You'd lose a vendor, not the system.

Not while we're managing it. Deployment and onboarding are done for you, and day-to-day use doesn't require an engineer on staff. If you have an IT person or firm you already trust, we work with them.

You can keep using them; Eddy doesn't require ripping anything out. Eddy's focus is the organization around the model: a shared memory built from your meetings, email, calendar, documents, and systems, with permissions that follow each person, actions you can require approval for, and a deployment you own. The model answering is your choice, and you can change it over time.

Some do, and it's improving. With those products, the memory lives in the vendor's service under the vendor's terms. With Eddy, the memory of your people, projects, and decisions lives in a system you control, whichever model is answering.

They share some traits: self-hosted, open source, and able to take action. OpenClaw is designed as a personal assistant that an individual installs and runs. Eddy is built for an organization: shared memory with access controls, actions that can require a person's approval, and meetings and phones included. We deploy and run it for you, with our responsibilities set out in the agreement.

Most organizations start on a dedicated server we operate for them: quick to set up, with nothing for you to maintain. When it makes sense, the same system can move to hardware you own, sized with your IT team. Running on your own hardware from day one is also possible. Whether it's worth the upfront cost depends on your usage and team size, and we'll give you an honest assessment during the assessment.

For much of everyday business work, such as summarizing, drafting, and running routine workflows, current open models do well. Where a task calls for a frontier model, Eddy can use one, under that provider's terms.

Eddy can send routine background work to smaller models and save larger ones for the tasks that need them. Actual costs depend on your volume and model choices, and we estimate them during the assessment.

Every engagement follows the same three stages: a Transformation Assessment, Implementation in separately scoped stages, and Ongoing Managed Care. Scope and pricing for each stage are agreed up front, and detailed terms are in each proposal. As a rough guide, the assessment fee typically runs about 10% of the deployment cost, though every engagement is quoted individually. See how we work.

Tell us about your organization. We read every inquiry ourselves and reply within two business days. If it looks like a fit, we'll send an invitation to talk.