Capture and query the reasoning behind changes made by AI agents in a codebase, providing a 'git blame' for the 'why'.
Sponsored
Selvedge provides a crucial long-term memory solution for codebases primarily developed by AI agents, acting as a 'git blame' that surfaces the *intent* behind code modifications. Instead of inferring reasoning from diffs, it captures the AI agent's live context and rationale at the moment each change is made. This ensures that even months or years later, developers can understand the 'why' behind schema updates, dependency shifts, or function alterations, preventing the loss of critical decision-making context that typically evaporates when agent sessions end.
Key Features
01Captures AI agent reasoning live and in context, not inferred post-hoc.
02Provides a robust CLI for blaming, diffing, history, and full-text search.
03Tracks changes at an entity level (DB columns, tables, functions, API routes, env vars, dependencies).
04Groups related events into named 'changesets' for comprehensive feature or task tracking.
054 GitHub stars
06Ensures data integrity with atomic schema migrations and concurrent write support for agent pools.
Use Cases
01Auditing and understanding the history of schema decisions, migrations, and dependency changes in long-term AI-coded codebases.
02Helping solo developers using AI remember the intent behind their AI-built project changes.
03Debugging and maintaining codebases where the original AI agent session context or developer's memory has evaporated.
