NotFair: goal-driven autonomous marketing agents that optimize business metrics via disciplined local loops

What it solves

NotFair provides autonomous, goal-driven marketing agents that manage business metrics (like organic clicks or ad spend) without requiring constant manual dashboard monitoring. It eliminates the need for "babysitting" marketing platforms by turning plain-language ambitions into measured, server-verified metrics and executing a disciplined improvement loop to reach those targets.

How it works

Users state a goal in plain language (e.g., "Grow organic clicks 30% in 90 days"). NotFair provisions a dedicated agent for that goal which verifies the metric via connected data sources and establishes a baseline. Once the user confirms the plan, the agent enters a loop: it measures the metric mechanically, reviews past moves against predicted effects, and executes at most one new move per check.

It runs locally on the user's machine using existing LLM harnesses like Claude Code or Codex, storing data in a local SQLite database. It connects to marketing platforms (Google Ads, Meta Ads, X Ads, Search Console, etc.) via MCP servers and OAuth, and can even execute code changes via GitHub pull requests.

Who it’s for

Business owners and marketers who want to autonomous automation of their SEO and paid ad campaigns without managing API keys or hosted runtimes.

Highlights

• Goal-centric interface: One screen per goal with a progress chart, check diary, and chat for steering.
• Disciplined autonomy: Agents operate within a set spend envelope and follow observation-window rules to prevent erratic behavior.
• Server-verified metrics: Metrics are re-run server-side to ensure agents are judged on reproducible numbers, not self-reported data.
• Local-first architecture: Runs on the user's machine using their own LLM subscriptions (Claude Code/Codex) and local SQLite storage.
• Extensive marketing skills: Includes a comprehensive plugin for manual audits, copywriting, and technical SEO work.