Distribution is one of the least structured parts of building a product. GrowthClaw approaches it as a state machine with explicit transitions, evaluation gates, and periodic strategy review.
Instead of prompting an agent to “do marketing,” it models the business, generates tasks, and runs them through a controlled pipeline.
Key pieces:
* Context intake from website + structured questions
* Strategy generation with explicit KPIs and constraints
* Task backlog stored in SQLite
* Human approval gate before execution
* Per-task evaluator that returns PASS / REVISE / ESCALATE
* A PM agent that enforces deterministic state transitions
* Strategy evolution that runs on cron three times per day
Everything runs locally. Workflows are YAML-defined. No external services required.
Curious whether treating distribution as infrastructure resonates with others here.
Distribution is one of the least structured parts of building a product. GrowthClaw approaches it as a state machine with explicit transitions, evaluation gates, and periodic strategy review.
Instead of prompting an agent to “do marketing,” it models the business, generates tasks, and runs them through a controlled pipeline.
Key pieces: * Context intake from website + structured questions * Strategy generation with explicit KPIs and constraints * Task backlog stored in SQLite * Human approval gate before execution * Per-task evaluator that returns PASS / REVISE / ESCALATE * A PM agent that enforces deterministic state transitions * Strategy evolution that runs on cron three times per day
Everything runs locally. Workflows are YAML-defined. No external services required.
Curious whether treating distribution as infrastructure resonates with others here.
more slop spam, here's my question to you
"If you can do this with Ai so easily, why do I want to use yours instead of the one my Ai generates?"
project is less than 24hrs old, 4 commits if I'm being generous
https://github.com/mrrkrieg/growthclaw/commits/main/