Repository map
Aider
Aider selects a compact, task-relevant map of the repository for the active model.
Aider codebase context combines repository maps, selected files, Git history, conventions, and model reasoning. Graphify adds a persistent relationship graph for dependencies, call paths, documentation, and architecture, with an installer that writes Graphify guidance into AGENTS.md. Use both: Aider's repository map is a live context-selection aid, while Graphify is a reusable structural artifact the assistant can query across sessions.
Published Updated
Graphify stores inspectable artifacts so codebase knowledge can be reviewed by humans and reused by AI coding assistants.

Repository map
Aider selects a compact, task-relevant map of the repository for the active model.
Persistent map
Graphify stores dependencies, calls, communities, and documentation relationships in graph artifacts.
Workflow
The Aider installer adds Graphify commands and verification policy to AGENTS.md.
Aider codebase context is the set of files, repository-map information, Git history, conventions, and model messages Aider uses during a coding session. The context window is finite, so choosing the right structural information matters as much as the model's raw coding ability. A persistent graph helps the assistant decide which files and relationships belong in the active working set. The useful unit is a bounded relationship answer that makes the next source read more targeted.
Aider's repository map is generated dynamically to fit the current task and token budget. Graphify is a persistent project artifact that captures explicit relationships, confidence labels, communities, and documentation links for repeated queries and human review. Neither map proves runtime behavior; their value is in reducing discovery cost before source and tests are examined.
Run graphify aider install after installing graphifyy. The installer writes a Graphify section to AGENTS.md, where Aider can find project-specific commands, graph paths, and rules for querying before broad repository reads. The installer is not a replacement for the graph build, and it does not ask Aider to ignore its own repository map.
Ask for the smallest structural answer that supports the task: callers, affected nodes, a path between two symbols, or a community summary. Then add the relevant source files to Aider so the model can reason from exact code instead of graph summaries alone. Bounded queries protect the context window and produce a clearer audit trail than asking for the whole project. If a result is surprising, inspect the graph edge in source before adding more files.
AGENTS.md should contain working rules rather than a copy of the graph. List the real validation commands, protected paths, graph artifacts, refresh procedure, and the requirement to verify graph-derived claims before changing code. The goal is a short operating contract that remains stable when individual files move.
Refresh after interface changes, module moves, dependency upgrades, or generated-code changes. Use update mode for code-only structure and a full extraction when docs, PDFs, diagrams, or other semantic inputs changed; then confirm the graph matches the current commit. In long-running branches, rebuild after merges that change interfaces instead of trusting the graph from the branch point.
graphify aider install
graphify extract .
graphify affected "AuthService" --depth 2Yes. The current Graphify CLI includes graphify aider install and writes its workflow guidance to AGENTS.md.
No. Aider's map is useful for live context selection; Graphify adds a persistent structural artifact.
Yes. Aider can work alongside Graphify CLI queries and local graph artifacts without an MCP server.
No. A graph can identify likely affected paths, but type checks and tests determine whether a change works.
Continue through the Graphify topic cluster with pages that explain assistant context, architecture review, documentation, and graph-based codebase understanding.
Persistent codebase memory for Claude Code
Give Claude Code durable codebase memory: build a Graphify graph, keep GRAPH_REPORT.md and graph.json in the project, and refresh them as the code changes.
Codex codebase context without rereading the repository
Give Codex persistent codebase context with graph artifacts, concise AGENTS.md rules, and MCP tools so repository questions use structure, not repeated scans.
AGENTS.md vs a codebase knowledge graph
Compare AGENTS.md instructions with a codebase knowledge graph, see what each layer stores, and learn how to combine them for persistent AI coding context.
Start from the use cases page or upload a project ZIP to build a graph that your team can inspect and reuse.