Issue Station
Sign In
Tools hereToolbox

hermes - 💡(How to fix) Fix feat: lcm — self-hosted long-term context memory plugin [1 participants]

Official PRs (…)
ON THIS PAGE

Recommended Tools

×6

Utilities matched from this issue’s tags and category — try them while you read without losing context.

GitHub issue graph ai analysis

Paste a GitHub issue URL. We fetch that issue, discover linked issues from bodies/comments/timeline, collect linked pull requests, and produce a structured English report.

The report is written in English Markdown for sharing and archival.

Helpful · Quick feedback

Loading…
GitHub stats
NousResearch/hermes-agent#15570•Fetched 2026-04-26 05:26:32
View on GitHub
Comments
0
Participants
1
Timeline
4
Reactions
0
Author
xyzmr114xyzmr114
Participants
xyzmr114xyzmr114
Timeline (top)
labeled ×4
RAW_BUFFERClick to expand / collapse

Motivation

Hermes needs a self-hosted, zero-vendor-lock-in memory plugin that is useful out of the box without requiring API keys for embeddings.

Current options:

  • Mem0: requires API key, hosted service
  • Honcho: good but not self-hostable for embeddings
  • Builtin memory: manual only, no auto-capture

We want: session auto-capture with LLM fact extraction + hybrid search, using any embedding provider.

Proposed Solution

lcm — a new built-in memory plugin:

  • Pluggable embedding providers: Ollama (local/free, default), OpenAI, Azure OpenAI, AWS Bedrock
  • Session auto-capture: buffers turn pairs, runs extraction every N turns, deduplicates similar facts
  • Hybrid BM25 + vector search with cross-encoder reranking
  • Memory tools: lcm_store, lcm_recall, lcm_list, lcm_forget, lcm_stats
  • CLI: hermes memory lcm start|stop|status|logs|setup

PR

See PR #15569 — includes the full plugin implementation.

Status

  • Plugin implementation (lcm MemoryProvider)
  • Bundled Docker service with pluggable embeddings
  • CLI commands
  • CI passing
  • Code review

Provider Support

ProviderStatus
Ollamadefault
OpenAIsupported
Azure OpenAIsupported
AWS Bedrocksupported
Google Vertex AIfuture

extent analysis

TL;DR

Implement the proposed lcm memory plugin to achieve self-hosted, zero-vendor-lock-in memory functionality with session auto-capture and hybrid search.

Guidance

  • Review the PR #15569 implementation to ensure it meets the requirements and is free of bugs.
  • Test the lcm plugin with different embedding providers (Ollama, OpenAI, Azure OpenAI, AWS Bedrock) to verify its pluggability.
  • Focus on getting the CI passing and completing the code review to move the implementation forward.
  • Consider adding more providers like Google Vertex AI in the future to increase the plugin's versatility.

Example

No code snippet is provided as the issue does not contain specific code that needs modification or implementation.

Notes

The success of this implementation depends on the quality of the PR #15569 and the thoroughness of the testing and review process.

Recommendation

Apply the proposed lcm memory plugin workaround to achieve the desired functionality, as it addresses the current limitations of existing memory plugins.

Vote matrix · Quick signals

Works
Did the solution work? Tap to confirm.
Easy Fix
Was it a quick fix?
Time Saver
Did it save you time?
Blocking
Was it severely blocking?
Common Issue
Are others likely hitting this too?
Flaky / Intermittent
Is it intermittent?
Verified / Reproducible
Can you reproduce it reliably?
Loading…
#api#ssr#installation#tensor shape#autograd error

Still need to ship something?

×6

Another batch ranked right after the header list — different links, same matching logic.

Back to top recommendations

TRENDING