PMB

Stop re-explaining your project to AI coding agents

Developer ToolsArtificial IntelligenceGitHubOpen Source

▲ 183 votes49 commentsLaunched Jun 29, 2026

Daily #1Weekly #5Monthly #2163

PMB gives Claude Code, Cursor, Codex and Zed persistent project memory through MCP. It stores decisions, lessons, goals, recent work, project facts and docs in one SQLite workspace on your disk. No cloud, no API keys, no LLM call on the read path. It is open source, offline-first, inspectable/exportable, with a local dashboard and honest impact tracking so you can see which memories actually help.

AI Analysis

📝 Summary

PMB is an open-source, offline-first tool that equips AI coding agents like Claude, Cursor, Codex, and Zed with persistent project memory via MCP. It stores decisions, lessons, goals, recent work, facts, and docs in a local SQLite file on disk. Core features include no cloud dependency, no API keys or LLM calls on reads, inspectable/exportable data, a local dashboard, and impact tracking to evaluate memory usefulness. It solves the major pain point of repeatedly re-explaining project context to AI tools, which wastes time and reduces accuracy. The value proposition is private, transparent, cost-free project continuity that boosts developer productivity and lets users own their data completely.

📈 Market Timing

In 2025-2026, AI coding agents are seeing explosive adoption as developers integrate them into daily workflows, yet context window limitations and memory loss remain critical bottlenecks. Trends favor local-first and privacy-centric AI tools amid growing concerns over data sovereignty and API costs. With maturing MCP standards and demand for autonomous agents, PMB addresses a clear gap at the perfect moment. Excellent Timing.

✅ Feasibility

High. Technical implementation uses mature SQLite and MCP integrations with low complexity. Operational costs are negligible as it runs fully locally with no servers or cloud bills. Open-source model enables community contributions, minimizing team burden. No significant compliance or supply chain risks due to offline design. Excellent scalability for individual to team use on standard hardware.

🎯 Target Market

Main segments: Software developers, full-stack engineers, and indie hackers (ages 25-45) heavily using AI coding tools on long-term projects. Industries: Tech, software development, open-source. Geographic: Global with concentration in US, Europe, China, and India. TAM for AI developer tools exceeds $10B by 2026; SAM for context/memory solutions approx. $500M-$1B; SOM for local/open-source tools in the tens of millions. Core pains: context loss and repetitive prompting. High willingness to pay for productivity gains, though PMB's open-source nature suggests reliance on sponsorships and optional paid add-ons.

⚔️ Competition

Medium. Direct competitors: 1. Continue (continue.dev), 2. Aider (aider.chat), 3. Mem0 (mem0.ai), 4. Claude Projects (anthropic.com/claude), 5. Cursor Memory features (cursor.com). PMB advantages: completely local SQLite with zero cloud/API costs, fully inspectable/exportable memory, built-in honest impact tracking, and true offline-first design. Disadvantages: narrower out-of-box integrations than commercial tools, depends on MCP adoption, and lower brand visibility as a newer open-source project. Strong differentiation through privacy, transparency, and no recurring costs.

Upgrade Pro to unlock full AI analysis