Public API · Early access developer preview
LynkMesh

Give your AI coding agent a map of your codebase.

LynkMesh analyzes your repository and generates structured, deterministic context so AI coding agents can understand dependencies, architecture, and relevant code before they start working.

Built for developers building AI coding agents.

Deterministic analysis — no LLM inference Structured Context Pack — mesh_context_ai_pack.v0.1 Live public API — api.lynkmesh.it.com
The problem

Your AI agent can write code. But does it understand your repository?

Large codebases are more than files and tokens. Dependencies, call relationships, structure, and boundaries matter.

The solution

LynkMesh turns your repository into AI-ready context.

LynkMesh extracts deterministic structural facts from your repository and packages them into a compact Context Pack that an AI agent can consume.

Live API Result

The API is live. This is what it returns.

One repository analysis produces a structured Context Pack containing repository facts that can be consumed by an AI coding system. No LLM inference runs inside the analysis.

Real responses from api.lynkmesh.it.com

The health endpoint is open and needs no API key — open it in your browser. The Context Pack below is one real analysis response from the compact profile.

One real run on a small PHP repository. Not a benchmark — pack size and content vary with repository structure.

GET /v1/health · live · no API key
{
  "status": "ok",
  "service": "lynkmesh-api",
  "api_version": "v1"
}
Real API response · Context Pack (compact)
{
  "job_id": "d5b3d3751fd34375b4bcc57b69201f33",
  "status": "done",
  "generated_at": "2026-09-14T14:30:28Z",
  "profile": "compact",
  "content_hash": "62bb5ad881ce…",
  "pack": {
    "schema_version": "mesh_context_ai_pack.v0.1",
    "executive_context": { "node_count": 7, "edge_count": 7 },
    "context_budget": {
      "estimated_input_tokens": 575,
      "estimator": "deterministic_char_heuristic_v0.1"
    },
    "guardrails": {
      "deterministic_facts_only": true,
      "contains_llm_inference": false,
      "must_cite_evidence_ids": true
    }
  }
}
What you get

A repository intelligence layer for your agent.

Four properties make the Context Pack useful as machine input for AI coding systems.

Built for AI agent builders

Don't build repository intelligence from scratch.

Building an AI coding agent means solving more than generation. Your agent also needs to navigate repositories, discover dependencies, and establish structural context.

LynkMesh provides a repository intelligence layer that your agent can call through an API.

Developer API

One API call. Structured repository context.

Submit a GitHub repository URL. Poll the job. Retrieve the Context Pack. Authentication is a single X-API-Key header.

    api.lynkmesh.it.com
    # POST https://api.lynkmesh.it.com/v1/analyze
    # header: X-API-Key: <your-key>
    {
      "repo_url": "https://github.com/your-org/your-repo",
      "branch": "main",
      "profile": "compact"
    }
    
    # → 202 Accepted
    {
      "job_id": "1f3c9a…",
      "status": "queued"
    }
    
    # → poll, then retrieve
    # GET /v1/jobs/{job_id}
    # GET /v1/jobs/{job_id}/context  → Context Pack
    Evidence & trust

    Don't take our word for it.

    LynkMesh is open about how it works. Explore the source code, generated Context Packs, and validation artifacts.

    01 / Baseline

    Evaluation method

    The baseline explains how to compare AI coding workflows without LynkMesh versus with LynkMesh, while keeping claims conservative and reviewable.

    • Repeatable before/after structure
    • Metric schema and wording rules
    • Public-safety guidelines
    • No benchmark or performance guarantee
    Open baseline →
    02 / First run

    Fixture-level evidence

    The first committed run uses a synthetic PHP mini shop fixture and includes deterministic CLI artifacts, before/after transcripts, screenshots, and a conservative comparison summary.

    • Expanded AI Context Pack artifact
    • Readable node and edge evidence labels
    • Guardrails against LLM inference claims
    • Human correctness scoring still treated as pending where applicable
    View run evidence →
    Early validation baseline Not benchmark proof No “AI became smarter” claim Deterministic graph evidence
    Under the hood: MeshContext

    The structured format behind the Context Pack.

    MeshContext is the structured format used to package repository intelligence for downstream AI systems. It keeps deterministic graph facts separate from interpretation, runs no LLM inference inside the analysis, and stays diff-ready. The sample below shows values from the public evidence fixture.

    mesh_context_ai_pack.v0.1
    {
      "schema_version": "mesh_context_ai_pack.v0.1",
      "profile": "compact",
      "purpose": "ai_consumption",
      "executive_context": {
        "node_count": 61,
        "edge_count": 99
      },
      "context_budget": {
        "estimated_input_tokens": 691,
        "estimator": "deterministic_char_heuristic_v0.1"
      },
      "guardrails": {
        "deterministic_facts_only": true,
        "contains_llm_inference": false,
        "must_cite_evidence_ids": true
      }
    }
    Local
    Coming soon

    Working directly with your local codebase?

    LynkMesh is also being developed as a local MCP integration for AI coding workflows, allowing developers to use repository intelligence locally without sending their source code to a remote service.

    Roadmap

    What's shipping now, next, and later.

    Principles

    No magic. Evidence first.

      Get started

      Give your agent a better context.

      Start with a repository. Get structured context your AI coding system can consume.

      Deterministic Code Intelligence AI-Ready Context Pack Structured Repository Context Submit → Poll → Retrieve No LLM Inference in Analysis Open-Core Engine Deterministic Code Intelligence AI-Ready Context Pack Structured Repository Context Submit → Poll → Retrieve No LLM Inference in Analysis Open-Core Engine