Selvedge Review 2026: Long-term memory for AI-coded codebases.

What Selvedge Does

Selvedge is a local MCP server that gives AI coding agents a durable memory of why code changed. As Claude Code, Cursor, or Copilot work, Selvedge logs the reasoning behind every edit — captured live, in the same context that produced it — and stores it in a SQLite file next to your code. Instead of guessing from a generated commit message months later, you can run selvedge blame payments.amount and see the original intent.

Key Features

• Live intent capture — Agents log change rationale in-context via log_change as they work.
• Entity attribution — Tracks users.email, env/STRIPE_SECRET_KEY, api/v1/checkout, deps/stripe.
• prior_attempts memory — Before editing, agents see what was tried and reverted — active memory.
• Editor auto-wiring — selvedge setup detects and wires supported editors automatically.
• Agent Trace export — Emits history in a portable observability format.

Pros

• Captures the ‘why’, not just the ‘what’
• Runs fully local with zero telemetry
• Works with Claude Code, Cursor, Copilot, Cline, Windsurf
• Prefix-searchable entity attribution
• Exports Agent Trace records for observability

Cons

• Requires agents to actively call its MCP tools
• No cloud sync across machines
• Young project, smaller community
• SQLite store needs manual backup
• Best value only on long-lived codebases

How Selvedge Compares

Selvedge is not alone. These tools also tackle similar problems:

• Git blame — Shows who/what changed, never the why.
• Code review tools — Review-time quality, not provenance.
• Agent observability (LangSmith etc.) — Trace LLM calls, not code intent.

Want a head-to-head? Read our Selvedge vs Cursor comparison.

Verdict

Selvedge earns a 4.2/5 (8.4/10). Long-term memory for AI-coded codebases. It is worth a look if you value agent orchestration and local-first workflows.