Personal use
Plan a garden, renovate a room, organize family work, or deliver for clients. CanDoItAll gives one person the structure and specialist support of a small virtual company.
The alpha now runs on Windows, macOS, Linux, and in Docker. Guided setup is available from the current main branch.
Useful at human scale
You do not need an enterprise program to benefit from structured projects and private agents. Begin with work you already understand.
Create a garden plan around the plants you want, the available space, sunlight, watering, season, budget, and work dependencies. Agents can help research options and turn the decisions into a layout, schedule, and task list. The same pattern works for renovations, trips, learning plans, or family events.
Give repeatable roles to research, planning, delivery, review, administration, or customer preparation. A freelancer remains accountable while a team of specialized agents helps keep several commercial projects organized inside one laptop.
A capable gaming PC can host a good local model, turning hardware bought for games into private home AI infrastructure. Family members can connect their own CanDoItAll instance to the local endpoint and keep sensitive conversations away from an external model provider.
A practical first loop
The value is not only in asking a model a question. It is in keeping the answer connected to the plan, files, responsibilities, and next action.
Create a project with the goal, constraints, useful files, and the decisions already made.
Ask an agent for options, risks, dependencies, and a realistic sequence; approve only the changes you want.
Use structure, Gantt, calendar, files, workflows, and assigned agents as views of the same project.
Retain outputs, approvals, status, and revised plans so the project becomes easier to continue later.
Private AI at home
Connect CanDoItAll to a local Ollama-compatible model endpoint and keep the model runtime on hardware you control.
Privacy still depends on the complete setup: keep storage, backups, plugins, remote tools, and any fallback providers inside the boundary you intend. A local model does not make an otherwise connected workflow automatically private.
Review provider setup
Costs for one person
Agent expenses do not have to disappear into one monthly provider bill. CanDoItAll keeps detailed usage and calculated cost records for conversations and for agent work inside workflows and processes.
A home user can compare paid services with local-model usage. A freelancer can see which agents, providers, and automated runs consume budget, then use those records when estimating work, reviewing margins, or explaining project costs.
See agent cost analytics
Use provider distribution and observed cost to decide when private home hardware is enough and when a paid model earns its place.
Use conversation, workflow, and process records as an input when estimating delivery, reviewing client profitability, or adjusting how an agent is used.
Repeated failures, retries, an expensive provider mix, or a runaway automated run become visible before they quietly consume the project budget.
More economy depth · planned later in 2026
Not every economy-related backend function is visible in the UI yet. Planned Ledger and Economy drivers will extend project budgeting, planning, simulation, and economic execution beyond today’s detailed usage and expense records.
Learn it your way
Everything can be installed and explored independently. A short paid session can help a family or freelancer choose suitable hardware, a private-model setup, and one useful first project without weeks of trial and error.
