Overview
pumaDB and Kickbacks.ai are two very different products aimed at developers working with AI agents, but they solve completely separate problems. pumaDB provides durable, lightweight memory storage for AI agents — allowing them to remember facts, preferences, project context, and state across sessions without setting up a database. Kickbacks.ai, on the other hand, is an ad marketplace that monetizes the wait states (spinner animations) in AI coding tools like Claude Code and Codex, sharing 50% of ad revenue with the developer.
Feature Comparison
| Feature | pumaDB | Kickbacks.ai |
|---|---|---|
| Core Purpose | Durable memory for AI agents | Monetize AI agent wait states |
| Target Users | Developers needing persistent agent context | Developers using Claude Code/Codex |
| Integration | MCP server or REST API | VS Code/IDE extension or CLI |
| Data Storage | JSON rows (20 tables, 1,000 rows each, 25 MB) | None |
| Memory Types | Skills, preferences, task state, research, etc. | Not applicable |
| Safety Features | Version history, rate limits, natural edits | Not applicable |
| Revenue Model | Subscription (free tier available) | Ad revenue sharing (50% to developer) |
| Setup Complexity | Minimal — add server URL or API key | Minimal — install extension |
Pricing
pumaDB offers a free tier that includes 20 tables, 1,000 rows per table, and 25 MB total storage. This is sufficient for many small to medium agent projects. Paid plans likely provide higher limits, but specific pricing details are not fully public on the website.
Kickbacks.ai is free for developers to use. Revenue is generated when advertisers bid for sponsored status lines that appear during agent wait states. Developers receive 50% of the ad spend. Advertisers bid per 1,000 impressions (five-second intervals), with clicks billed at 50x the impression rate.
Pros and Cons
pumaDB Pros
- No database setup or infrastructure management required
- Supports both MCP (for agents) and REST (for server-side apps)
- Built-in version history and natural language editing
- Lightweight schema designed specifically for agent memory
- Rate limits and safety rails prevent abuse
pumaDB Cons
- Storage limits may be restrictive for large-scale applications (25 MB total)
- Pricing for higher tiers is not clearly documented
- Requires internet connectivity for hosted service
Kickbacks.ai Pros
- Passive income from AI agent idle time
- Simple setup — just install an extension
- Transparent revenue sharing (50% to developer)
- Works with popular AI coding tools (Claude Code, Codex)
Kickbacks.ai Cons
- Only works during agent wait states (spinner animations)
- Revenue potential is uncertain and likely low per user
- Limited to VS Code/IDE extensions and CLI
- No memory or productivity features — purely monetization
Verdict
pumaDB and Kickbacks.ai serve entirely different purposes. pumaDB is a practical tool for developers who need persistent agent memory without database overhead — ideal for building smarter, context-aware AI workflows. Kickbacks.ai is a novel monetization experiment for developers who want to earn small revenue from their AI coding tools' idle time. Choose pumaDB if you need memory; choose Kickbacks.ai if you want to experiment with passive income.
Kickbacks.ai
pumaDB