Connection
Base URL, credentials and a reachable server.
AI coding workspace
Source symbol search
Find the relevant occurrence before editing it. SourceLens brings source search, declarations and call sites into a project conversation you can follow.
● ● ●SourceLens●
A concrete starting point · Source symbol search ×
rg -n 'TargetSymbol' src tests \
--glob '!generated/**'
# Inspect declarations and callers before changing the symbol.✳Search for a symbol in source and tests before editing it.
✓ ChangesReview before applying ↗
Illustrative workspace
Project
Exclude generated files and separate declarations, real calls and incidental text matches.
Review before applying
Compare actual behavior with the expected result. If a check fails, investigate it before applying more changes.
A concrete starting point
rg -n 'TargetSymbol' src tests \
--glob '!generated/**'
# Inspect declarations and callers before changing the symbol.Turn source matches into an informed editing decision
A search result identifies text, but it does not explain which occurrence controls the behavior you need to change. SourceLens is a Windows AI coding workspace for examining that distinction. Select the repository folder and ask a connected model to relate the relevant matches to the task before proposing an implementation.
Consider a symbol that appears in source, tests and generated output. Use the project conversation to establish where it is declared, which callers matter and whether generated files should be edited through their source instead. The search example illustrates a possible investigation; it is not a claim that SourceLens includes a specialized indexing engine.
Once the responsible path is clear, request a focused change with an explanation of the affected callers. Preserve enough surrounding context to assess why the edit belongs there. Review the diff and run the project checks associated with those paths, especially when a matching name appears in more than one module with different responsibilities.
Model selection remains configurable through supported OpenAI-compatible or Anthropic endpoints. If your endpoint offers an uncensored AI model, its actual behavior and tool support still depend on the model and server. Use those capabilities to support the investigation, while keeping evidence from your repository at the center of the decision.
Turn source matches into an informed editing decision — practical questions
Should I send every matching file to the model?
Begin with the declaration and the callers related to the reported behavior. Add more files when the investigation exposes a dependency. A large collection of unexplained matches can obscure the distinction between relevant implementation and incidental text.
How should generated files affect the proposed patch?
Identify the generator or source input before editing generated output. Ask the model to explain the regeneration path if it is visible in the repository. Review the resulting source change and regenerate using your established project process.
What makes a search-based explanation convincing?
It should connect an occurrence to the execution path or contract in question. Ask for the relevant caller and the behavior it influences. Matching a name alone does not establish that changing it will solve the reported issue.
Can SourceLens guarantee complete repository coverage?
Do not assume complete coverage from a conversation or search command. Excluded paths, generated sources and external consumers can matter. Context limits and available tools depend on your configured endpoint; state any unexamined areas when reviewing the proposal.
SourceLens · Updated 2026-10-25
