Skip to main content

BUG: UI hangs when running mcp_add_memory_tctnzi tool

Bug Reports

OS: macOS 26.4.1

Version: BoltAI 2.12.1 (build 68) - Setapp

LLM: Anthropic Claude Sonnet 4.6

I instructed the LLM to commit something in particular to memory.

It appears to have called the mem0-memory-mcp server installed. Although it’s not clear to me if that’s what Bolt tried to use, or whether it’s using the built-in Bolt memory tool. I can see in the UI what you see in attached screenshot.

The conversation hang after it tried to use this memory tool. I had to manually terminate the hung prompt response.

I can see that the number of memories in the Bolt Memory tool didn’t increase. So if that is the memory system it’s utilising in this chat, the hung attempt to add a memory did not get so far as actually adding a memory.

Posted by Inspired Earth ·
Status: Completed3 comments

Log in to comment and vote

Comments3

  • Daniel Nguyen changed status to In QA
    Team•

    Jul 19

    Pinned

    I’ve fixed this in v2.14. Can you confirm?

  • Inspired Earth

    •

    May 13

    I might have figured out the issue:

    Previous config (hanging)

    • Command: uvx mem0-mcp-server — meaning every startup, uvx had to ephemerally fetch, resolve, and bootstrap the package from the uv cache before the server could respond. This startup latency was almost certainly what caused BoltAI to hang—it was timing out waiting for the MCP server to become ready.

    • User ID: jonathan — an entity that had no memories stored against it, so even if it connected it would have returned nothing.

    New config (working)

    • Command: /Users/jonathan/.local/bin/mem0-mcp-server — a persistently installed binary (via uv tool install mem0-mcp-server), so it starts immediately with no resolution or download step.

    • User ID: mem0-mcp-user — the correct entity where existing memories are actually stored.

    The core issue for BoltAI was almost certainly the uvx ephemeral execution model—it's fine for one-off CLI tools but too slow to initialise for MCP servers that need to respond to a handshake promptly on startup.

  • Inspired Earth

    •

    May 13

    I asked the LLM which memory tool/MCP server it was trying to use. It came back with:

    Good question — and I should be transparent about what happened. I was attempting to use the memory tools that are available in this conversation, specifically the mcp_add_memory_tctnzi function which appears to be a Mem0-based MCP tool (based on the function naming pattern). That's the one that was available to me in this session, not a BoltAI-native memory system.

    It looks like the call either timed out or failed silently. Let me try again now to store the two things worth remembering from this conversation:

    Tool: mcp_add_memory_tctnzi

    Running tool: mcp_add_memory_tctnzi...

    It hung again. So it seems for now I can’t use the mem0 MCP server in BoltAI.