Show HN: PlanWright – A control plane for AI coding agents
4 hours ago
- #AI Agents
- #Software Development
- #Compliance
- Planwright addresses the shift in software development bottlenecks from coding speed to human-paced ceremonies like planning and acceptance, now that coding agents are faster.
- It inverts planning by using AI to synthesize unstructured inputs (like meeting transcripts and Slack) into structured objectives with machine-checkable criteria, where humans approve instead of write.
- It inverts acceptance by triaging completed objectives—automatically clearing mechanical criteria and routing only judgment-based items to humans, with cryptographic signatures for audit trails.
- The system integrates with Claude Desktop and Claude Cowork, using a board to manage objectives, agent execution via MCP, and human approvals, creating hash-chained records for compliance.
- Pricing tiers include Free for solo developers, Team for small teams, Business for compliance needs, and Enterprise for large organizations with advanced features.