log-mcp

log-mcp is a Python-based MCP server that analyzes large log files for AI agents. It provides tools to scan log overviews, search by regex/level/time, extract segments, deduplicate and summarize errors, compute log statistics, compare patterns across files, and classify lines using an optional Rust TF-IDF classifier plus an optional BERT-mini re-scoring stage. It can run with or without the Rust classifier (Python fallback).

Evaluated Mar 30, 2026 (22d ago)
Repo ↗ DevTools mcp log-analysis observability ai-ml rust-inference python tf-idf bert error-deduplication local-processing
⚙ Agent Friendliness
51
/ 100
Can an agent use this?
🔒 Security
40
/ 100
Is it safe for agents?
⚡ Reliability
14
/ 100
Does it work consistently?

Score Breakdown

⚙ Agent Friendliness

MCP Quality
78
Documentation
55
Error Messages
0
Auth Simplicity
95
Rate Limits
10

🔒 Security

TLS Enforcement
80
Auth Strength
25
Scope Granularity
10
Dep. Hygiene
55
Secret Handling
40

Likely runs locally as a local MCP server (no auth described). README includes an external installation command and references handling of an ANTHROPIC_API_KEY for training/labeling, but does not document secret storage practices or logging redaction. Network/TLS/auth for the MCP transport are not described; dependency hygiene cannot be verified from provided content.

⚡ Reliability

Uptime/SLA
0
Version Stability
35
Breaking Changes
0
Error Recovery
20
AF Security Reliability

Best When

You have large (tens of MB to many GB) logs where you want an AI agent to narrow down interesting lines/patterns before sending summaries back into a context window.

Avoid When

You need a standards-based web API with documented HTTP error codes, strict auth/scope semantics, or guaranteed consistent ordering/results; or you must run in environments without permission to execute local commands/models (it appears to run locally via MCP server).

Use Cases

  • Analyzing large log files efficiently for root-cause style insights (errors, anomalies, security events).
  • Generating grouped/deduplicated error summaries with stack traces/fingerprints.
  • Comparing two or more runs to find patterns unique to each file or frequency outliers.
  • Surfacing anomalies that may not use explicit ERROR/FATAL levels via classification.
  • Building agent workflows that need structured log querying (overview/search/extract/analyze).

Not For

  • Interactive, manual browsing of raw log contents (the tools compress output for AI).
  • Use as a general-purpose log search UI without agent integration.
  • Situations requiring strongly validated data contracts, strict pagination semantics, or guaranteed deterministic outputs (not evidenced).

Interface

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

Authentication

Methods: Local execution via MCP CLI integration (no described user auth).
OAuth: No Scopes: No

README describes installing/running the MCP server locally and registering it with Claude; no authentication or scope model is described for the MCP transport.

Pricing

Free tier: No
Requires CC: No

No pricing model for the package itself is described; README references potential Claude Batch API costs for labeling.

Agent Metadata

Pagination
none
Idempotent
False
Retry Guidance
Not documented

Known Gotchas

  • No explicit tool I/O schema, argument validation rules, or MCP error taxonomy is provided in the README content; agents may need to handle unexpected tool failures themselves.
  • Optional ML stages (Rust TF-IDF, optional BERT-mini with GPU/Metal) may cause variability in runtime/latency and failure modes depending on environment.
  • Classifier can be used where explicit ERROR/FATAL is missing, so agents should not assume level-based results are complete.
  • Large-file processing may be slow depending on hardware and classifier availability; agent timeouts/retries are not documented.

Alternatives

Full Evaluation Report

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

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