Selvedge

Long-term memory for AI-coded codebases — including what was already tried and rejected.

Line attribution tells you who wrote something. Selvedge tells your agent whatnotto write next: the approaches this codebase already tried, reverted, and why. It's agit blamefor AI agents, for thewhyrather than which model touched which line — captured live, by the agent, as the change happens, so nothing downstream has to guess at it.

Selvedge is a local MCP server. AI coding agents (Claude Code, Cursor, Copilot) call it as they work to log structured change events with reasoning. Your data stays in a SQLite file under.selvedge/next to your code.

Local-first by default, team-server by choice, zero-LLM always.

Six months ago, your AI agent added a column calleduser_tier_v2. You don't know why.git blamepoints to a commit fromclaude-codewith a generated message that says "Update schema." The session that made the change is long gone — and so is the prompt that produced it.

$ selvedge blame user_tier_v2 user_tier_v2 Changed 2025-10-14 09:31:02 Agent claude-code Commit 3e7a991 Reasoning User asked to add a grandfathering flag for legacy free-tier users during the pricing migration. Stores the original tier so we can backfill discounts without touching billing history.

That reasoning wascaptured by the agent in the moment— written into Selvedge from the same context that produced the change. Not inferred from the diff afterward by a second LLM. Not a hand-typed commit message.

Selvedge has two audiences. Same tool, samepip install, same SQLite file under.selvedge/. Different scale of pain.

Teams running long-term, AI-coded codebases.When the project is big enough that you (or someone else) will touch it again in six months, twelve months, three years — but most of it was written by an agent whose context evaporated the day each PR shipped.git blametells you what changed. Selvedge tells youwhy— even after the agent session, the prompt template, the developer who asked for it, and the model version are all long gone. This is the original use case: production codebases, schema decisions, migrations, dependency changes that need an audit trail that survives turnover.

Solo developers using Claude Code on everyday projects.Side projects, weekend builds, the small internal tool you keep poking at. You don't need enterprise governance — you just need to remember why you (or your agent) did the thing you did yesterday, last week, last sprint. Runselvedge initonce. Add four lines to yourCLAUDE.md. From then on,selvedge blameis muscle memory — a way to talk to your past self when your past self was an LLM.

If you've ever come back to your own AI-built project and thought "what was thisforagain?", Selvedge is the missing piece.

Human-written code leaks intent everywhere — commit messages, PR descriptions, inline comments, the Slack thread that preceded it. AI-written code doesn't. The agent has perfect clarity about why it made each decision, but that context lives in the prompt and evaporates when the conversation ends.

Six months later, your team is debugging a schema decision with no trail.git blametells youwhatchanged andwhen. It can't tell youwhy.

Selvedge captures the why — live, by the agent itself, as the change is made.The diff is git's job. The why is Selvedge's.

The memory comes to the agent, and the store gets its dials.Two themes, shipped together because the config half is what the rest needed to read settings from.

Delivery.Selvedge already blocked re-edits of reverted entities. What was missing was delivery when there is nothing to veto. Two new hooks:

- SessionStartinjects a compact digest as a session begins — decisions due for a revisit, entities that were tried and reverted, recent changesets.
- PreCompactfires just before context compaction destroys this session's reasoning and names the watched entities you edited but never logged.

Both are quiet when they have nothing to say, size-capped, read-only, and templated. Neither can block anything — PreCompact deliberately declines the veto the hook API offers it. This is the answer to a measured failure mode: two 2026 papers recorded pull-model memory tools going unused entirely (zero voluntary memory operations across 114 turns against a pre-seeded store) while deterministic injection landed every time.

selvedge export --format markdownrenders the store as a reviewable digest to commit next to it, so captured intent shows up in a pull request instead of hiding inside a binary. Deterministic — regenerating with no new events is a zero-line diff.

Config..selvedge/config.tomlis now first-class, with a canonical precedence chain thatselvedge doctorprints per setting. It brings:

- selvedge prune --include-events— the first path that can delete captured reasoning, so it needsbotha confirmation andSELVEDGE_DESTRUCTIVE=1. Neither alone is enough, because--yesin a cron entry defeats a prompt and a shell profile defeats an env var. Events retention defaults to never.
- Event-size bounds(diff_bytes,reasoning_bytes) that truncate loudly — a marker in the text, a warning at write time, a count inselvedge stats.
- Secret-shape warningsatlog_change, extendable viaredaction_patterns, plus adoctorrow that scans what's already stored. Warn, never reject.

Also:five review issues closed. The enforcement hook's allow path is40% faster(33.6 ms → 20.1 ms per gated call) andSELVEDGE_HOOK_DISABLE=1finally short-circuits before the imports it was documented to skip;log_changeno longer discardsrevisit_after/constraint/stale_whenon renames and supersedes; the CLI's--jsonand the MCP tools now return identical structures; and the Docker image no longer ships the maintainer's own database. Tests 826 → 984.

Fixes a broken install, and lands a full code-quality pass.mcp2.0.0 (released 2026-07-28) removedmcp.server.fastmcp, and Selvedge declaredmcp>=1.0.0with no upper bound — so anypip install selvedgeafter that date pulled 2.0.0 andselvedge-serverfailed at import. This release pins the dependency.If your server stopped starting, this is why — upgrade.

It ships alongside a review that put nine independent passes over the codebase and then tried todisproveevery finding before acting on it. Seventeen confirmed defects fixed. The ones you would actually have noticed:

- The enforcement hook stopped blocking things it shouldn't.Reading a tracked file —cat,git diff,pytest,ruff check— was blocked, and the remediation the error message told you to run was blocked by the same gate, so there was no way out from the CLI. Two more paths fed the same false blocks: a commented-out line of SQL counted as a real deletion, and any commit message merely containing the word "revert" marked every file it touched as reverted.
- Lookups got fast at scale.The main entity read was scanning every row — measured 7.4 ms → 0.35 ms at 100k events, and the hook had been taking seconds on large stores.
- selvedge setupcan no longer delete parts of yourCLAUDE.md, an interrupted backup can no longer destroy your last good one, and upgrading while two Selvedge processes are running no longer crashes with an error that looked like database corruption.

Tests went 739 → 826. No schema change and no tool-surface change, so this isdrop-in for anyone on 0.3.9.x.

AI agents call Selvedge as they work. Selvedge captures thewhyinto a durable, queryable store and emits it back out — asAgent Tracerecords for cross-tool readers, as observability metadata that links into Sentry/Datadog stack traces, and as compliance artifacts for SOC 2 and EU AI Act audits.

Why "rejected paths" matter — the one that isn't copyable.The expensive failure isn't forgetting why a column exists. It's an agent confidently re-implementing something the team already killed for a good reason, six months after everyone who knew that left the context window. None of the line-attribution tools above surface rejected paths at all, and it isn't a feature gap they can close in a release — a line-oriented store has no notion of an entity that persisted across a try → revert → retry cycle. Seedocs/demos/prior-attempts.md.

Why determinism matters.Selvedge's reasoning is the agent's own intent, written from the same context window that produced the change. There is no model anywhere in the storage or retrieval path, so the same query returns the same answer today and in two years, across model versions. Tools that infer reasoning post-hoc are running a second LLM that never saw the original prompt: what it produces is paraphrase, and re-running it can produce different categories for the same change. As a Hacker News commenter put it about a competing approach,"grep won't find your commit because you rejected 'oauth-library'… unless there is deterministic enforcement"(0x457).

Selvedge ↔ Agent Trace.Agent Traceis an open AI code-attribution wire format published by Cursor (RFC, Jan 2026). Its original GitHub home went 404 in August 2026 and the multi-vendor momentum behind it has faded, but the spec and schema still resolve at agent-trace.dev, frozen at v0.1.0. Sincev0.3.9,selvedge export --format agent-traceemits Agent Trace v0.1.0 records andselvedge import --format agent-tracereads them back — a portable, documented interchange format for file/line AI attribution, with reasoning and entity-level provenance carried in each record'sdev.selvedgemetadata. The mapping is indocs/agent-trace-interop.md; Selvedge vendors the schema and has no runtime dependency on the upstream project.

/plugin marketplace add masondelan/selvedge /plugin install selvedge@selvedge

pip install selvedge cd your-project selvedge setup

selvedge status # recent activity + missing-commit count selvedge diff users # all changes to the users table selvedge diff users.email # changes to a specific column selvedge blame payments.amount # what changed last and why selvedge history --since 30d # last 30 days of changes selvedge history --since 15m # last 15 minutes ('m' = minutes) selvedge changeset add-stripe-billing # all events for a feature/task selvedge search "stripe" # full-text search selvedge stats # log_change coverage report (per-agent) selvedge import migrations/ # backfill from migration files selvedge export --format csv # dump history to CSV

Selvedge is a standard stdio MCP server, so it works with any MCP client — Claude Code, Cursor, Windsurf, Codex CLI, Gemini CLI, and more. SeeWorks with any MCP clientfor the exact config per client. For Claude Code:

claude mcp add selvedge -- selvedge-server

Prefer to copy-paste? The same block is one click away on the website:selvedge.sh/prompt-block— with a copy button and notes on what your agent does with it.

claude mcp add selvedge -- selvedge-server

{ "mcpServers": { "selvedge": { "command": "selvedge-server" } } }

Docs:https://code.claude.com/docs/en/mcp

{ "mcpServers": { "selvedge": { "command": "selvedge-server" } } }

Cursor's newer schema also accepts an explicit"type": "stdio"; thecommand-only form works too (Cursor infers stdio fromcommand). Docs:https://cursor.com/docs/mcp

{ "mcpServers": { "selvedge": { "command": "selvedge-server" } } }

Windsurf hot-reloads the file — no restart needed. The in-appPlugins → View raw configbutton opens the exact file Cascade reads. Docs:https://docs.windsurf.com/windsurf/cascade/mcp

[mcp_servers.selvedge] command = "selvedge-server"

Or runcodex mcp add selvedge -- selvedge-server. Docs:https://developers.openai.com/codex/config-reference

{ "mcpServers": { "selvedge": { "command": "selvedge-server" } } }

Or rungemini mcp add -s user selvedge selvedge-server. Docs:https://github.com/google-gemini/gemini-cli/blob/main/docs/tools/mcp-server.md

{ "mcpServers": { "selvedge": { "command": "selvedge-server" } } }

The diff is git's job. Thewhyis Selvedge's.