MCP servers that ship as code — a repository or a package you install and run locally (usually over stdio). They have no hosted endpoint, so we cannot probe or grade them: instead each one shows its repository, install command, license, stars and when it was last updated.
Collected from GitHub repositories and npm, PyPI, Cargo, Go, Maven, NuGet and RubyGems packages tagged as MCP servers, the official MCP Registry, Docker's MCP catalog, ModelScope, Baidu MCP World, Cline and other public directories; one entry per repository. A server that also offers a hosted endpoint is listed under Remote instead. These entries are not counted in our reachability, verification or readiness statistics.
-
MCP server over Microsoft Graph: four read-only audit evidence tools any AI client can call
MIT -
aging-knowledge-graphgithub.com/GiorgioHuang/aging-knowledge-graphupdated this year
An open, evidence-traceable knowledge graph for healthy aging — Postgres + pgvector, REST + MCP API, natural-language search, and Claude-powered curator/reviewer agents.
MIT -
AndyAI Knowledge Factory with ASAL: structural awareness, repo graph, AST pipeline, MCP graph tools, hybrid RAG+Graph routing and evidence-backed impact maps.
MIT -
Agent-agnostic, MCP-native OSINT hub: 25+ native collectors, evidence graph, web console — one curl from a fresh Linux VM to a running stack.
★ 1 -
Evidence-first cybersecurity MCP server with persistent SQLite graph memory, vulnerability triage, scoped testing and reports for Antigravity and Codex.
MIT ★ 1 -
Local, versioned research evidence graph for Mudcrab, with source citations, material/native/GPU observations and read-only MCP.
Apache-2.0 -
Tauri 2 desktop intelligence platform — Cesium.js 3D globe, 185 live panels, MCP server for Claude Code, and an explainable algorithm intelligence layer (truth scoring, evidence graph, compound risk, forecast calibration) plus shortage / weather / insights engines. 600+ deterministic tests.
NOASSERTION ★ 12 -
Local MCP server with a code graph across all repositories of a project: symbols, calls, and cross-repo HTTP / config / package links with file:line evidence.
npm Apache-2.0 -
Five messy data sources fused into one evidence graph, and an AI agent that investigates it through MCP tools, with every answer backed by evidence ids
MIT -
Evidence-first entity and knowledge graph MCP server for AI agents to verify claims, trace cryptographic provenance, and retrieve 90% less context.
MIT ★ 1 -
A deterministic code graph for codebases: every caller, route, page and table a change touches, with file:line evidence. CLI, MCP server and web view for AI agents and reviewers.
pypi MIT ★ 4 -
Deterministic code intelligence for repositories: local dependency graphs, impact analysis, and bounded MCP evidence.
oci Apache-2.0 -
Evidence-backed static code graph intelligence for MQL5 repositories, with CLI, dashboard, reference corpus, and an experimental Codex MCP plugin.
MIT ★ 2 -
Incident root-cause agents on a FalkorDB temporal graph: MCP graph tools, edge-cited evidence, time-aware queries, and human-confirmed memory. Graph Hacks 2026, Tracks 01 + 02.
MIT -
The memory database for AI agents - graph memory, hybrid search, tamper-evident thought/edge journal.
pypi MIT ★ 12 -
Evidence-first code intelligence and repository knowledge graphs for AI coding agents via MCP.
npm Apache-2.0 -
TypeScript knowledge graph RAG for AI agents: source evidence, hybrid retrieval, MCP integration, and a permission-aware PostgreSQL runtime.
Apache-2.0 ★ 31 -
llm-sca-toolinggithub.com/grammy-jiang/llm-sca-toolingupdated this year
Evidence-first, LLM-augmented static code analysis: graph indexing, MCP server, audit workflows
pypi MIT -
AI agents that find money-laundering rings no single bank can see, on FalkorDB: one graph per bank, a consortium graph, echo-proof evidence, signed reports and a hosted MCP server.
MIT ★ 1 -
Local code architecture graph for Python and TypeScript: trace events, contracts, dependencies, and change impact with source evidence.
MIT -
factorforgegithub.com/aman9824/factorforgeupdated this year
Vectorless retrieval (no embeddings) over an evidence-linked knowledge graph, served as a standalone MCP server. Multi-agent Claude research pipeline.
MIT -
Persistent, evidence-backed code graph for coding agents. CLI and MCP server.
cargo Apache-2.0 ★ 3 -
Agent-native incident investigation graph: analysts and AI agents (MCP/REST) link evidence from EDR, Entra ID, cloud and GitHub logs into one shared, real-time attack graph and timeline. Self-hosted, no database.
★ 5 -
Native Weavatrix SEO CLI and MCP host: Search Evidence Graph audits, plans, and retrieval for code-driven websites.
npm MIT -
Built by AI, for AI: a knowledge graph that agents use as long-term memory, where every fact cites its evidence. MCP and REST, one Rust binary.
Apache-2.0 ★ 1 -
Blast Radius — Forge app for Jira, Rovo and MCP: Teamwork Graph–powered release risk verdicts, evidence, checklists and suggested reviewers (Codegeist 2026)
MIT -
Behavioral diff for AI-written code. Your agent changed 40 files — Veris tells it what behavior actually changed and whether any of it was verified. MCP-native, checker-resolved call graph, tamper-evident evidence. Never executes your code. Local-first, MIT.
npm MIT ★ 2 -
ravengithub.com/DigiBugCat/ravenupdated this year
raven — epistemological memory for AI agents. Knowledge graph with derivation chains, gated writes, and confidence earned from evidence. Python + Postgres/pgvector MCP service; successor to pantainos-memory.
-
@1337xcode/cortexgithub.com/1337xcode/cortexupdated this year
Code intelligence MCP server. Indexes your repository into a confidence-aware call graph with SCIP ingestion, framework adapters, and evidence-fusion ranking for AI coding agents.
npm MIT -
Evidence-linked repository change-impact graph with a minimal MCP stdio interface. Unknowns stated as unknown.
MIT -
Give your coding agent a map: local repository context, code graphs, stateful navigation and JSON-LD evidence through 18 stdio MCP tools.
NOASSERTION -
context-sourcegithub.com/peterica/context-sourceupdated this year
TypeScript static analysis → Entity/Relationship/Evidence graph in SQLite, explored via Web UI and queried by AI through MCP.
MIT -
Local codebase intelligence CLI + MCP server for AI coding agents: SQLite code graph, 28 languages, 287 commands, 246 MCP tools, change-safety gates, audit evidence, zero API keys.
go Apache-2.0 ★ 516 -
Quant-Agent-OS: open-source quantitative intelligence platform for AI clients (MCP, SDK). Quant Research (Graph RAG with cited evidence) and Quant App Factory. Pre-alpha.
NOASSERTION -
Local-first code graph and change impact analysis for AI coding agents. MCP server, SQLite, tree-sitter. Python and TypeScript. Every edge carries evidence and a confidence label.
pypi Apache-2.0 ★ 1 -
Local-first memory, knowledge and evidence infrastructure for AI agents. Persistent state, hybrid search, graph + MCP. Signed releases, one-line install. Free tier.
NOASSERTION ★ 2 -
Turn AI and tech news into market signals with evidence: MCP server, knowledge graph, time-series signal store, swappable LLMs. Local first.
NOASSERTION -
Distill AI coding sessions (Claude Code / Codex / Cursor) into weekly reports, ADRs, incident reviews and a knowledge graph — local-first, evidence-linked.
pypi MIT ★ 4 -
vic3-graphraggithub.com/realsxb/vic3-graphragupdated this year
Auditable Victoria 3 GraphRAG mapping player slang to official mechanics with concept graphs, dual-view retrieval, MCP tools, and traceable evidence.
MIT -
SENGRA is an autonomous fraud investigation agent that uncovers organized fraud rings through multi-hop graph analysis. It leverages WCC, PageRank, Louvain communities, and betweenness centrality, retains entity intelligence across sessions, and visualizes evidence paths behind every decision in a real-time dark ops console.
-
Approval-gated MCP server for Microsoft Entra ID + Intune via Graph: reads, human-approved writes, fail-closed admin guard, redacted audit, evidence-backed acceptance matrix
MIT -
architecture-intelligence-platformgithub.com/michaelegner/architecture-intelligence-platformactive
Trusted architecture context for AI agents: an evidence-backed architecture graph from OpenAPI, AsyncAPI and OpenTelemetry, served over MCP.
Apache-2.0 ★ 10 -
osint-investigatorgithub.com/allsmog/osint-investigatorupdated this year
Target-agnostic OSINT investigation MCP server and plugin for evidence ledgers, archiving, verification, entity graphs, and source analysis.
-
Open-source AI context engine for Unity projects: GitNexus code graph, Unity GUID/fileID scene and prefab graph, Roslyn symbols, runtime evidence, and read-only MCP/Context API for coding agents.
Apache-2.0 ★ 2 -
🔎 Username, email & phone OSINT — 9,700+ lookup checks from 19 projects, with profile enrichment, bulk searches, evidence graphs and expert-reviewed updates.
pypi GPL-3.0 ★ 272 -
Open source, self-hosted semantic layer and knowledge graph for AI agents. Evidence-backed definitions, human review, and reusable MCP context.
Apache-2.0 -
From a PubMed search to a knowledge graph of the literature: full texts from PubMed Central, the evidence around key papers and an Obsidian knowledge graph, from the command line, Python or an AI assistant. Includes a study of the retracted Aβ*56 paper.
pypi NOASSERTION -
MCP server and one-shot CLI for Outlook and Exchange mail over Microsoft Graph: evidence-bearing search, a read-only multi-mailbox plugin, and sender and recipient allowlists for agents.
MIT ★ 3 -
MCP server and structured reference for AI law: obligations, regulations, provisions, evidence, and enforcement deadlines across jurisdictions. The obligation-registry layer of the PAICE legal graph.
npm SEE LICENSE IN DATA-LICENSE.md -
Self-hosted observability for agentic debugging: traces, logs, replay, AI cost, Agent Action Graphs, MCP, profiles, and usage analytics in one evidence graph.
MIT ★ 2 -
Agentic Version Control System — an AI-native VCS where intent, session, operation, evidence and decision are first-class objects and code is a materialized projection of an operation graph.
npm Apache-2.0 -
Protect customer promises when Jira plans change. RippleGuard connects commitments to Jira work, flags delivery risks, and explains the evidence through Rovo, Teamwork Graph, and read-only MCP tools. Built on Atlassian Forge for Codegeist 2026.
-
Local-first personal knowledge management (PKM) MCP server and learning memory for every AI tutor—goal trees, a reusable knowledge graph, and evidence-gated mastery.
mcpb Apache-2.0 -
Read-only MCP access to a BAOS workspace: product graphs, requirements, evidence, research.
npm -
Local MCP server that indexes code into a graph and vector store and answers where it is, who calls it and what breaks - with file:line evidence, fully offline.
NOASSERTION ★ 1 -
Index Python, Java, TypeScript and JavaScript into an embedded code graph and answer callers, inheritance, dead-code and change-impact queries with file:line evidence. CLI, MCP server and a labelled retrieval evaluation.
MIT -
ssot-registrygithub.com/groupsum/ssot-registryupdated this year
A single source of truth for software delivery. Use the Python CLI to link features, specs, and ADRs to claims, tests, and evidence, certify releases, and export lineage graphs.
pypi Apache-2.0 ★ 5 -
Local-first project map for AI-assisted programming: scan JS/TS into a feature topology with line-level evidence, expose it to AI agents over MCP, and verify changes against the graph.
MIT -
Evidence-backed context graph and MCP server for AI coding agents
Apache-2.0 -
A live architecture map of your codebase, on screen beside Claude Code and in the notes your agent reads while it works. Evidence-backed graph with file and line proof, served over MCP and a CLI.
MIT ★ 8