silex

Local

silex.tec.brDeveloper Tools0Updated 15w ago

About silex

silex is a local-first MCP server that reconstructs code architecture, dependencies, workflows, and business rules for coding agents. Its findings include verified file-and-line evidence, and it flags what cannot be established instead of guessing.

Key features

01Codebase architecture and workflow mapping
02Verified file-and-line evidence
03Confidence and provenance for findings
04Detection of unproven areas
05Shareable evidence reports

Use cases

01Understand a codebase before modifying it
02Trace dependencies and business rules
03Check evidence behind an agent's codebase claims

MCP support

Exposes reconstructed codebase understanding to AI coding agents through MCP. It runs locally and supports stdio connections from MCP clients.

Configuration

Local server config is the same for every client. Copy it into your client’s MCP config file.

mcp.json · Universal config

{
  "mcpServers": {
    "silex-tec-silex": {
      "command": "uvx",
      "args": [
        "silex-archaeology"
      ]
    }
  }
}

Run commanduvx silex-archaeology

Local servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.

Packages

pypisilex-archaeology0.11.38stdio

Tools

Saved tool list · updated 3w ago

Parameter	Type	Required	Description
use_cache	boolean	No	serves the cached contract (``.silex/workspace.db``) when the source content for the scope has not changed — content-hash key ``sha256(business_logic.json)`` + ``package_id``. Default ``True``; ``False`` forces generation. Automatically invalidates when the source file changes (never serves stale data).
package_id	string	No	handle determinístico do escopo (``pkg_l{level}_{sha256[:12]}``). Use este OU ``community_id``.
community_id	string	No	UUID da community em ``code_communities`` (alternativa a ``package_id``; ``package_id`` tem preferência se ambos vierem).
workspace_id	string	Yes	workspace UUID (mesmo de ``generate_domain_jsons``).
artifact_root	string	No	raiz alternativa de ``.silex/domain`` (default: ``.silex/domain``). Útil para staging/testes.

Details

From the project README

Silex — evidence-based code archaeology

Understand the codebase before changing the code.

Silex reconstructs architecture, dependencies, workflows and business rules from your source code — and serves that understanding to your AI coding agents via MCP. Every claim is anchored to file:line with a verified quote. What Silex can't prove, it tells you.

Local-first. Your code never leaves your machine. No account, no card, no telemetry.

→ Site & docs: https://silex.tec.br · Concepts · MCP tool reference · llms.txt

Install (free tier)

# Python 3.10+
pip install silex-archaeology

Until the PyPI release lands you can install the wheel from Releases: pip install https://github.com/silex-tec/silex/releases/latest/download/silex_archaeology-py3-none-any.whl

Connect it to your MCP client (Claude Code, Cursor, Claude Desktop, Windsurf, Cline, Continue, Zed) — stdio config:

{
  "mcpServers": {
    "silex": { "command": "silex-mcp-server", "args": ["serve", "--stdio"] }
  }
}

Full install guide (Windows/PowerShell notes included): https://silex.tec.br/docs/

Try it on a real codebase

Clone Microsoft's eShop (MIT) and point a scan at it — the Concepts page shows the community graph, the recovered order state machine and anchored findings from that exact scan, so you can compare your output against a published reference.

What you get (free)

Local MCP server, full codebase scan (C#, JS/TS, Java, PHP, COBOL, VB…)
Community map — the real shape of the system, vendor/generated classified out of the way by default
Evidence anchored to file:line, confidence with provenance
Unknown-territory detection: what can't be proven is flagged, not guessed
Shareable evidence report
Works with the AI seat you already pay for

Pro tier (team workspaces, deep architecture visualization, refactoring trust contracts): coming soon — join the waitlist.

Issues & feedback

Bug reports and field feedback are welcome in Issues. The engine is proprietary; this repository hosts distribution, documentation and the public roadmap conversation.

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

Read the getting-started guide

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.