← Blog
April 26, 2026

MentisDB 0.9.5.41 — MCP 2025-11-25, Hermes Native Integration, Dashboard Fixes

0.9.5.41 upgrades MentisDB's MCP protocol support to the latest specification and marks the first third-party agent to ship a native MentisDB memory plugin. Hermes, the open-source agent from Nous Research, now connects directly to mentisdb for durable semantic memory without needing an MCP bridge.

Hermes now has a native MentisDB MemoryProvider. The Hermes team implemented their own memory plugin against the MentisDB API for direct agent-level integration — no intermediary MCP bridge, no extra process, just Hermes talking directly to a running mentisdb instance.

MCP 2025-11-25 Protocol Support

The cloudllm_mcp dependency has been upgraded from 0.1.0 to 0.2.1, bringing full support for the MCP 2025-11-25 protocol revision. The initialize handshake now negotiates "2025-11-25" instead of "2025-06-18".

Streamable HTTP servers now validate Origin headers for security on bound addresses, and transparently skip origin validation on unspecified addresses (0.0.0.0, ::, ::1) for compatibility with tools and SDKs that don't send Origin headers during the MCP handshake.

ComponentBeforeAfter
MCP Protocol Version2025-06-182025-11-25
cloudllm_mcp0.1.00.2.1
Origin ValidationAlways enforcedConditional (bound vs unspecified addr)

Hermes Native Integration

Nous Research's Hermes agent is an open-source AI agent framework designed for research, reasoning, and autonomous task execution. The Hermes team built a native MemoryProvider plugin that speaks the MentisDB protocol directly:

  1. Hermes reads initialize.instructions from the MCP handshake
  2. The instructions direct Hermes to load the MentisDB operating skill from resources/read(mentisdb://skill/core)
  3. Hermes bootstraps a memory chain and begins writing durable Checkpoints, Decisions, LessonLearned records, and Retrieval calls through the MentisDB MCP tools — the same way any other MCP-capable client does

This is the first community-driven native MentisDB integration, validating the vision of an open, agent-agnostic memory layer. Any agent framework can now follow the same pattern: implement a memory plugin against the standard MCP tools MentisDB exposes, and your agent gets durable semantic memory that survives context resets, model swaps, and session boundaries.

Dashboard Fix

The "+Bootstrap New Chain" and "↺ Refresh" buttons on the dashboard Chains page stopped working after a refactor in the previous release. The root cause was a one-shot _chainBootstrapWired guard that attached click handlers only once, but renderChainList() recreates the button DOM elements on every call. After any page navigation (Chains → Agents → back to Chains), the buttons rendered but did nothing.

The fix re-attaches the header button listeners on every render call while keeping the one-shot guard for the static modal elements that survive DOM replacement. This is safe and follows the same pattern used by every other dynamic table in the dashboard.

Upgrade

cargo install mentisdb --force

No migration required. Existing chains, skills, and agent registries are fully compatible. The MCP protocol version bump is negotiated transparently during the initialize handshake. If you use Hermes, make sure you're running the latest Hermes release with the native MentisDB MemoryProvider.