Skip to content

docs: competitive deep-dive — codegraph vs narsil-mcp#262

Merged
carlos-alm merged 2 commits intomainfrom
docs/competitive-narsil
Mar 3, 2026
Merged

docs: competitive deep-dive — codegraph vs narsil-mcp#262
carlos-alm merged 2 commits intomainfrom
docs/competitive-narsil

Conversation

@carlos-alm
Copy link
Copy Markdown
Contributor

Summary

  • Adds generated/competitive/narsil-mcp.md — a comprehensive competitive deep-dive comparing codegraph against narsil-mcp (postrv/narsil-mcp), our closest head-to-head competitor
  • Follows the same structure as the Joern deep-dive (joern.md in PR docs: competitive deep-dive vs Joern #260)
  • narsil-mcp is ranked Bump actions/checkout from 4 to 6 #2 in our competitive analysis (score 4.5, tied with Joern)

Document Structure

  1. Executive Summary — key dimensions table (mission, target user, graph model, core question, rebuild model, runtime)
  2. Problem Alignment with FOUNDATION.md — all 8 principles evaluated and scored (Codegraph 6, narsil-mcp 0, Tie 2)
  3. Feature-by-Feature Comparison — 9 sections, 130+ features:
    • A. Parsing & Language Support (24 features)
    • B. Graph Model & Analysis Depth (18 features)
    • C. Query Language & Interface (12 features)
    • D. Performance & Resource Usage (14 features)
    • E. Installation & Deployment (10 features)
    • F. AI Agent & MCP Integration (12 features)
    • G. Security Analysis (12 features)
    • H. Developer Productivity Features (18 features)
    • I. Ecosystem & Community (11 features)
  4. Where Each Tool is the Better Choice — choose codegraph / choose narsil-mcp / use both
  5. Gap Analysis — 6 features worth adopting, 8 not worth adopting (mapped to FOUNDATION.md)
  6. Competitive Positioning Statement
  7. Key Metrics Summary table

Key Findings

  • Codegraph wins on: persistent incremental builds, confidence-scored edges, compound CLI commands (35+), embeddable JS API, CI gates, impact analysis, code health governance, token efficiency (3,600 vs 12,000 tokens)
  • narsil-mcp wins on: language count (32 vs 11), security analysis (taint + OWASP + CWE + SBOM), neural search, CFG/DFG analysis, SPARQL/RDF, browser WASM, interactive visualization
  • Key insight: narsil-mcp maximizes breadth (90 tools); codegraph maximizes depth-per-tool and always-current guarantees

Test plan

  • Document follows joern.md structure exactly
  • All 8 FOUNDATION.md principles evaluated
  • 130+ features across 9 sections
  • Review factual claims against narsil-mcp GitHub README
  • Verify internal links

Loading
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant