# Koragraph > Koragraph turns every repository you point it at into one live code graph, served over MCP, so your AI agent reads your systems instead of rediscovering them. Koragraph turns a company's repositories into a living, evidence-backed code knowledge graph, so AI coding agents and the engineers working alongside them stop rediscovering the same codebase on every request. ## Product Koragraph MCP is the first thing Koragraph ships. One live graph of every repository you point it at, served over the Model Context Protocol, so the agent you already use reads your systems instead of rediscovering them. It is completely free and it runs locally: the graph is built and served from the developer's own machine, not a hosted service. Status: available now. Source and releases: https://github.com/Koragraph/KoragraphMCP Speaks the Model Context Protocol, so it connects to any MCP-capable coding tool. Confirmed compatible: Claude Code, Cursor, Windsurf, Cline, Continue, Codex, OpenCode, Zed, Antigravity, GitHub Copilot. ## Install - Install the server: npm install -g koragraphmcp - Ingest a repository: koragraph ingest /path/to/repo - Check the install: koragraph doctor - Add it to Claude Code: claude mcp add koragraph -s user -- koragraph mcp - Requires Node.js 22 or newer. Full walkthrough: https://www.koragraph.in/integrations/claude-code ## Tools this server exposes - explore: Symbol or plain English. Ranked declarations plus source, callers, callees. - blast_radius: Run before editing. What depends on this, and what has no test coverage. - search_code: Locate a name in the graph, not raw text. - neighbours: Callers and callees of one symbol. - changes_with: What has historically changed together with a symbol. - file_symbols: Declarations in one file. - overview: Orient on the first turn. Store level index across repos. - recall: A failed attempt and its fix, a hazard, a revert. - remember: Save a durable, code anchored fact. The only writer. Languages parsed, through tree-sitter: C, C++, C#, Go, Java, JavaScript, PHP, Python, Ruby, Rust, Swift, TypeScript and TSX. ## How to read this site - Landing page: https://www.koragraph.in - Benchmarks index, every head to head we have run: https://www.koragraph.in/benchmarks - Research index, papers with released task suites: https://www.koragraph.in/research - Korablog, notes on agent research: https://www.koragraph.in/blog - Glossary, the terms defined: https://www.koragraph.in/glossary - Learning hub, forty concepts from first principles drawn as one graph: https://www.koragraph.in/learn - Install guide for Claude Code: https://www.koragraph.in/integrations/claude-code - Full text of everything on this site, in one file: https://www.koragraph.in/llms-full.txt - Sitemap: https://www.koragraph.in/sitemap.xml Every claim below is measured and published with its inputs. ## Benchmarks ### Koragraph vs Graphify Declaration extraction, indexing speed and retrieval, across ten languages and 102 repositories. August 2026. Systems compared: Koragraph, Graphify v0.9.28 at depth 6. Koragraph indexes 99.70% of the declarations in 102 repositories against Graphify’s 82.07%, does it with 3.78× less CPU, and beats Graphify’s best retrieval result at any depth while spending 45% fewer tokens. - 99.70% (Graphify 82.07%): declaration recall across 10 languages - 3.78× (ahead on 17 of 17 repositories): less CPU to index - 0.824 (Graphify’s best 0.769, on 45% fewer tokens): retrieval recall against their ceiling - 4 of 6 (2 ties, none lost): context budgets won on retrieval Full results: https://www.koragraph.in/benchmarks/koragraph-vs-graphify Corpus, questions, gold sets and scorer: https://github.com/agilkatakam/graph-benchmark ## Research ### 94% of Frontier Retrieval Quality from a 494M-Parameter Model Method summarisation for code knowledge graphs in 2.43 GPU-hours, and the point where more training data stops paying. August 2026. Preprint. A 494M-parameter model fine-tuned for 2.43 T4-hours, about $0.85 of commodity GPU time, writes the method descriptions a code knowledge graph is searched on at 93.8% of the frontier teacher’s retrieval quality, and beats a hosted model Koragraph itself was paying for. Page: https://www.koragraph.in/research/small-model-method-summarisation PDF: https://www.koragraph.in/research/small-model-method-summarisation.pdf ### Can Graph-Based Context Engineering Substitute for Model Scale? Task Resolution and Cost for Cheap and Frontier Models on Repository-Level Code Tasks. August 2026. Preprint. A cheap model given graph-derived context solved 9 of 11 repository-level tasks at $0.027 per solved task. A frontier agentic system searching the same repositories with grep solved 9 as well, at $2.82. Page: https://www.koragraph.in/research/graph-context-vs-model-scale PDF: https://www.koragraph.in/research/graph-context-vs-model-scale.pdf Code: https://github.com/agilkatakam/Koragraph_test_suite_1 ## How the measurements are made - Systems are scored against an independent third-party referee, never against each other and never against either system's own definition of a correct answer. - For declaration extraction the referee is the language's own compiler front end wherever one can parse a single file without build configuration: go/parser, CPython ast, Roslyn, tsc, the Kotlin PSI. - For retrieval the questions are maintainers' own commit subjects and the gold answers are located by Universal Ctags, an indexer neither system uses. - Corpora are pinned to published commit SHAs. - Neither system spends a paid model call to produce its answers at measurement time. - Retrieval scoring is blind: the scorer sees no system identity, and rows are pooled and shuffled before grading. - Held-out corpora are cloned after development is finished and measured once, so corpus fitting is visible. ## Contact Source and releases: https://github.com/Koragraph/KoragraphMCP Email founders@koragraph.in to talk to us about a live repository. ## Notes for agents and their users - This file is generated from the same data the pages render from, so it cannot drift from what a human reader sees at https://www.koragraph.in. - Figures carry their comparison alongside them. Quote the comparison with the figure, since a number without the system it is measured against is not a result. - Scope matters and is stated on each benchmark page. Retrieval results cover four repositories of 20 to 63 source files with no cross-repository linking measured; extraction covers 102 repositories across 10 languages. Neither is a claim about codebases of arbitrary size. - Competitors are measured at their best documented setting, not their defaults. Graphify retrieval figures are at depth 6, its documented maximum. - Graphify wins pooled extraction precision, 99.86% against Koragraph 99.78%, and leads on C retrieval questions that do not involve preprocessor macros at two context budgets. - Koragraph MCP can be downloaded and run today from https://github.com/Koragraph/KoragraphMCP.