CODE / REPOSITORY UNDERSTANDING
THE SHORT ANSWER
A code knowledge graph represents a repository as connected files, modules, symbols, and code relationships so you can explore how the system fits together instead of reading only a file tree. Infinite Graph's repository-import workflow reads a local folder in the browser, builds a source-backed graph, and adds derived architectural views as interpretations for you to inspect—not as unquestionable facts.
Open Infinite GraphTHE PROBLEM
Directories show where files live, but the questions developers ask are usually relational: where does this value come from, which module calls this function, what depends on this package, and where does a change travel? A graph makes those paths visible while retaining the source locations that justify them.
A REVIEWABLE WORKFLOW
The product treats source facts, resolved relationships, and architectural interpretations as different kinds of evidence.
The browser-side reader selects source files from a folder, applies safety and size limits, and does not upload repository contents as part of acquisition.
The import pipeline records repository identity and provenance, then extracts supported files, modules, symbols, imports, exports, and other syntax facts.
Relationships such as imports, exports, calls, inheritance, and references can be represented with resolved, unresolved, or ambiguous status rather than hiding uncertainty.
Deterministic projection and structural or semantic overlays help compress a large repository into a usable view; derived interpretations remain reviewable and adoptable.
PRODUCT EVIDENCE
The code page is intentionally specific about the graph's evidence boundary and its staged availability.
IN PRACTICE
A developer imports a local TypeScript repository before making a billing change. They trace the path from a workspace control to an API handler and persistence boundary, inspect the source files behind the relationships, and use any architectural grouping as a prompt for investigation rather than a replacement for code review.
Illustrative workflow, not a measured onboarding or performance result.
A USEFUL DISTINCTION
A dependency graph usually answers which packages or files depend on which others. A code knowledge graph can include that view, but it also preserves symbols, source locations, repository history, provenance, and derived concepts that help explain the system to a human.
FURTHER READING
The architecture of a codebase and the architecture you need to understand it are not necessarily the same graph.
QUESTIONS WORTH ANSWERING
The browser acquisition layer reads a local folder and is designed not to upload or fetch repository contents during that step. The product still applies file, size, and safety limits before creating an import snapshot.
Coverage depends on the language adapters and import policy available in the deployment. Unsupported or failed files should be reported with diagnostics rather than presented as if they were analyzed successfully.
No. Source facts and supported static relationships have a different authority boundary from structural, semantic, or AI interpretations. Those derived views are there to help you explore and decide.
Not necessarily. Repository import is feature-gated, so availability depends on the product deployment and its configuration. This page explains the workflow without promising universal access.
WHEN YOU ARE READY
Open Infinite Graph in the browser and keep the decisions about your knowledge with you.
Open Infinite Graph