rlm-claude

Provides an MCP server for Claude Code that persists “infinite memory” across sessions by saving conversation chunks and extracting/saving reusable “insights”. It auto-snapshots before Claude Code’s /compact and supports search (regex, BM25, optional semantic embeddings), retention/archiving, and multi-project organization.

Evaluated Mar 30, 2026 (21d ago)
Homepage ↗ Repo ↗ Ai Ml mcp claude-code context-management memory search python self-hosted persistence
⚙ Agent Friendliness
62
/ 100
Can an agent use this?
🔒 Security
34
/ 100
Is it safe for agents?
⚡ Reliability
31
/ 100
Does it work consistently?

Score Breakdown

⚙ Agent Friendliness

MCP Quality
82
Documentation
78
Error Messages
0
Auth Simplicity
90
Rate Limits
10

🔒 Security

TLS Enforcement
20
Auth Strength
25
Scope Granularity
20
Dep. Hygiene
55
Secret Handling
60

Likely self-hosted/local MCP process with file-based storage. TLS/auth are not documented; if run remotely, transport security should be handled externally. No mention of encryption-at-rest for stored memory. The project claims safe I/O (atomic writes, path validation, locking) and uses structured local context directories, which is a positive sign for minimizing data corruption risks, but explicit secrecy handling and network exposure controls are not documented in the provided content.

⚡ Reliability

Uptime/SLA
0
Version Stability
40
Breaking Changes
45
Error Recovery
40
AF Security Reliability

Best When

You run Claude Code locally and want durable, searchable context using MCP tools with local persistence in ~/.claude/rlm/context/.

Avoid When

You cannot ensure local file permissions are safe, or you need strong documented guarantees around encryption-at-rest, auth, and operational recovery.

Use Cases

  • Persistent team/project memory for Claude Code across sessions
  • Searching prior decisions, preferences, and conversation context
  • Auto-saving relevant history before context compaction
  • Archiving/purging old conversation chunks with restore capability
  • Organization and retrieval of information by project/domain and chunk metadata

Not For

  • Security-sensitive data sharing over an untrusted network without transport/auth review
  • Fully managed/hosted memory services (this is self-hosted local storage)
  • Use cases requiring strict enterprise compliance guarantees that aren’t documented
  • High-concurrency workloads needing a hosted database backend instead of local files/vector storage

Interface

REST API
No
GraphQL
No
gRPC
No
MCP Server
Yes
SDK
No
Webhooks
No

Authentication

Methods: None documented (implied local/self-hosted MCP process)
OAuth: No Scopes: No

No explicit authentication mechanism is described in the provided README/manifest. Authentication expectations (e.g., local-only vs network-exposed MCP) are not documented.

Pricing

Free tier: No
Requires CC: No

Software is distributed under MIT; no subscription pricing is indicated.

Agent Metadata

Pagination
none
Idempotent
False
Retry Guidance
Not documented

Known Gotchas

  • Because authentication is not described, if the MCP server is reachable over a network, agents may need additional operational safeguards outside the package.
  • Retention/archiving/purge behavior affects recall results; agents should be careful to use restore/preview tools appropriately before relying on older context.
  • Semantic search depends on optional embedding dependencies/providers; results may degrade or differ when semantic deps are missing.
  • Hooks modify Claude Code configuration; ensure the user’s ~/.claude/settings.json merge/backup approach is appropriate for your environment.

Alternatives

Full Evaluation Report

Comprehensive deep-dive: security analysis, reliability audit, agent experience review, cost modeling, competitive positioning, and improvement roadmap for rlm-claude.

AI-powered analysis · PDF + markdown · Delivered within 30 minutes

$99

Package Brief

Quick verdict, integration guide, cost projections, gotchas with workarounds, and alternatives comparison.

Delivered within 10 minutes

$3

Score Monitoring

Get alerted when this package's AF, security, or reliability scores change significantly. Stay ahead of regressions.

Continuous monitoring

$3/mo

Scores are editorial opinions as of 2026-03-30.

8642
Packages Evaluated
17761
Need Evaluation
586
Need Re-evaluation
Community Powered