Ai

Claude-Mem v13.4.0 Adds OpenRouter Base URL Config and Clears Defect Backlog

Claude-Mem v13.4.0 Adds OpenRouter Base URL Config and Clears Defect Backlog

Claude-Mem's latest release, v13.4.0, is all about cleaning house and opening doors. The team tackled a mountain of defects — plans 01 through 11 plus standalone fixes — and slashed the failing test count from 46 to zero. Typecheck errors? Also gone, from 24 to nil. But the real headline is a new feature: you can now configure an OpenAI-compatible base URL for the OpenRouter provider. That means pointing claude-mem at DeepSeek, LM Studio, or any custom endpoint that speaks the OpenAI API. It’s a big deal for flexibility.

What Changed

The core change is the addition of the CLAUDE_MEM_OPENROUTER_BASE_URL environment variable. Previously, OpenRouter was locked to its default endpoint. Now users can route requests to alternative services. This isn’t just a cosmetic tweak — it’s a fundamental shift in how claude-mem can be deployed. For example, you could run it entirely locally with LM Studio, reducing latency and keeping data on-prem. Or you could swap in a cheaper provider like DeepSeek without rewriting integration logic.

Beyond that, the defect cleanup is extensive. The release notes list 11 specific fix plans and a slew of one-off corrections. While the changelog doesn’t detail each bug, the test suite results speak for themselves: 46 failing tests to zero. That’s a serious quality jump. The team also cleared 24 typecheck errors, which suggests improved type safety and fewer runtime surprises.

Why It Matters

This release makes claude-mem more adaptable and more reliable. For developers building on top of it, the OpenRouter base URL config is a game-changer. It decouples the project from a single provider, future-proofing it against API changes, pricing shifts, or outages. You want to test against a local model? Go ahead. Need to comply with strict data residency laws? Point it at a compliant endpoint. That’s real-world flexibility.

The defect backlog clearance, meanwhile, signals maturity. A tool that was once plagued by dozens of test failures is now solid. That’s the kind of stability that enterprises demand. And let’s be honest — seeing a project ship a version with zero test failures and zero typecheck errors is rare. It inspires confidence.

Personal take: This update feels like a turning point. Claude-Mem was already useful for AI memory management, but the provider extensibility moves it from a niche tool to a more universal building block. I’d be surprised if we don’t see more integrations popping up soon. The team clearly cares about quality — and that’s refreshing.

Official Source: https://github.com/thedotmack/claude-mem/releases/tag/v13.4.0