Shaperv1.0.0

A developer runs Claude Code inside a Shape Up team's repo and asks the agent to continue the current cycle. The agent calls get_active_work, reads the pitch spec via get_pitch_context, implements the scope, nudges the hill chart from 0.25 to 0.75, and marks the scope complete when tests pass.

A bootstrapped founder uses Shaper to run 6-week cycles alone and delegates implementation to an AI agent. The agent orients on the active cycle, picks the smallest unstarted scope, does the work, and reports hill progress without the founder opening the Shaper UI.

A product agency manages multiple Shape Up clients in Shaper and uses AI agents to onboard onto each cycle quickly. Agents read pitch documents for appetite and BDD scenarios, then execute scopes and update hill charts so account managers can report status without standups.

A designer writes pitches in Shaper and hands scopes to an AI agent to build UI components. The agent reads get_pitch_context for BDD scenarios and hill positions, then creates and completes scopes as the design spec is implemented.

An EM wires the Shaper MCP into a CI/agent harness so autonomous agents claim scopes from the active cycle. The EM uses get_active_work to see which scopes are over the hill and which agents are stalled, using positions as a live throughput signal.

Per-seat or per-workspace monthly subscription for the Shaper project management app itself, with API/MCP access bundled into paid tiers. This is the core revenue engine for a tool like Shaper.

Charge teams for MCP tool calls beyond a free tier, similar to LLM API pricing. Aligns revenue with how much agent work runs inside a workspace, appealing to heavy AI-adopting teams.

Annual contracts sold to companies that want Shaper embedded with SSO, audit logs, dedicated MCP endpoints, and support. Sold to engineering orgs running many Shape Up teams with agent fleets.