Make agent work add up
Practical guides to project context, reviewing results, managing costs and working together in VECTA.
Working with agents
Claude Code, Codex or Gemini CLI: which agent to use
Compare the work you need done, the tools each agent has and the result you can verify.
Give agents useful project context
Keep decisions, files and task conversations connected without pasting the whole project into every request.
How to review code written by an AI agent
Bring the conversation, changed files, checks and browser result into one review.
Hand over a task with its context
Give a teammate the goal, discussion, files and current decision instead of another disconnected chat.
Costs & time
How much does an AI coding agent cost per task?
Connect reported spending to individual tasks and projects instead of reading an unexplained total.
Set useful task budgets
Put cost, tokens and time next to the scope of a task before the work grows.
Review time alongside the result
Use elapsed time and task progress to understand where work is waiting or taking another pass.
Your workspace
Run an AI coding agent locally or in the cloud?
Choose where files and tools live separately from the model that helps you.
Continue an AI coding task from your phone
Use the browser to follow work running on a connected computer, with the same task conversation.
Understand VECTA’s access boundaries
Know who can see a task, where tools run and when an action needs your decision.