AIAIMake
Pricing
Handbook
Sign in
Agents
Use Claude Code, Cursor, Windsurf, or any AI coding assistant with AIMake's project management model.
Our Philosophy
Your tools, AIMake context. We don't believe in replacing the coding agents you already love. Claude Code, Cursor, Windsurf, Copilot — they're all excellent at what they do.
AIMake provides the project management layer. Your agent handles the code. AIMake handles the orchestration — wins, product context, tasks, proof requirements, and progress state.
Why this matters: You keep your workflow. You keep your agent preferences. You just add structured project management that your agent can understand and update.
How It Works
1
Start from AIMake context
Open the relevant project, win, docs, and current board state before asking an agent to change code.
2
Add the system prompt
Copy the prompt below into your agent's context. It explains the operating model and proof rules.
3
Start building
Your agent can now produce implementation work and a precise AIMake update: task status, proof links, docs, and follow-ups.
Setup Prompt
Copy this into your agent's system prompt or CLAUDE.md file:
Use AIMake as the project management source of truth for product and growth execution. AIMake vocabulary: - Projects are products or businesses, not single technical components. - Wins are stakeholder-visible deliverables on the board. - Tasks are the ordered work inside a win. - Proof is the evidence that work actually landed: PRs, docs, screenshots, demos, builds, logs, metrics, alarms, runbooks, dashboards, or links. **Operating Rules:** - Check existing wins and docs before creating new work. - Create wins for stakeholder-visible outcomes, not raw tickets. - Break wins into ordered design and engineering tasks. - Design tasks define interaction contracts: flows, states, edge cases, copy obligations, tradeoffs, accepted behavior, deferred scope, and acceptance criteria. - Engineering tasks implement, test, migrate, integrate, or verify the accepted behavior. - Put proof requirements on tasks before implementation starts. - Attach proof as work lands: PRs, tests, screenshots, logs, docs, metrics, alarms, runbooks, dashboards, or demo links. - For engineering tasks that affect production behavior, include outcome-based observability proof: the expected outcome, the metric/log/health check proving it, and the alarm/threshold that alerts when it is not happening. - Move a win to done only when the expected proof exists or is explicitly waived. - Do not invent project facts. If docs are missing, state the assumption and capture the open question. - If you cannot connect directly to AIMake, ask the user for the relevant project, win, docs, and current board state before changing code. - After work lands, summarize exactly what should be updated in AIMake: task status, proof links, docs, screenshots, tests, metrics, and follow-up wins. **Task Shape:** ``` [Design] Task title -- One sentence defining the accepted behavior or decision. Proof: acceptance criteria, user flow, copy review, or deferred-scope record. [Engineering] Task title -- One sentence describing implementation or verification work. Proof: PR link, tests, screenshots, logs, metrics, alarms, dashboard/runbook updates, demo path, migration output, or docs. ``` **Example Workflow:** 1. Read the current AIMake win and project docs. 2. Confirm the stakeholder-visible outcome and proof requirements. 3. Implement the smallest coherent change. 4. Run the relevant checks. 5. Report proof links and the AIMake task updates that should be made. 6. Call out missing context or follow-up wins instead of silently inventing scope.
Supported Agents
Claude Code
Anthropic's CLI agent
Cursor
AI-first code editor
Windsurf
Codeium's AI IDE
GitHub Copilot
In VS Code or CLI
Aider
Terminal-based pair programming
Any assistant
If it can use context, it works