High-performance code intelligence MCP server. Indexes codebases into a persistent knowledge graph — average repo in milliseconds. 158 languages, sub-ms queries, 99% fewer tokens. Single static binary, zero dependencies.
Drift inferred · capture-to-capture
No drift recorded — single capability capture; advisories appear once its surface changes.
tools
-
delete_project
Remove a project and all its graph data.
-
detect_changes
Map git diff to affected symbols + blast radius with risk classification.
-
get_architecture
Codebase overview: languages, packages, routes, hotspots, clusters, ADR.
-
get_code_snippet
Read source code for a function by qualified name.
-
get_graph_schema
Node/edge counts, relationship patterns, property definitions per label. Run this first.
-
index_repository
Index a repository into the graph. Auto-sync keeps it fresh after that.
-
index_status
Check indexing status of a project.
-
ingest_traces
Ingest runtime traces to validate HTTP_CALLS edges.
-
list_projects
List all indexed projects with node/edge counts.
-
manage_adr
CRUD for Architecture Decision Records.
-
query_graph
Execute Cypher-like graph queries (read-only).
-
search_code
Grep-like text search within indexed project files.
-
search_graph
Structured search by label, name pattern, file pattern, degree filters. Pagination via limit/offset.
-
trace_call_path
BFS traversal — who calls a function and what it calls. Depth 1-5.
last analysis: too-large
No code evidence — the analyzed source reached for no tracked permissions, tools, or hooks.