Open-source CLI · local-first · MCP server

Your AI agent learned the whole internet — but not your team's code.

Coding agents were trained on public repos, so they can't know the conventions your team settled in private PR reviews. difflore captures those team-specific rules and feeds them to Claude, Cursor, and Codex while they write — locally, no account.

Open-source runtime

Start locally with the CLI and MCP server. No cloud account needed.

Team cloud

Add hosted review-memory sync when private repos and shared policy matter.

cargo install --git https://github.com/difflore/difflore-cli difflore-cli

Apache-2.0 · runs locally · no account · works with 12+ agents

See how it works ↓View pricing
Imports fromGitHub PR reviewsAgentsClaude CodeOpenAI CodexCursorWindsurfZedGoose

How it works

A small approval loop, not another knowledge base.

difflore stays useful by keeping the memory set tight, sourced, and approved by the people who own the code.

  1. 1

    Import review history

    Start with merged PR comments where your team already said what mattered.

  2. 2

    Approve reusable rules

    Accept only patterns worth repeating; dismiss one-off comments before agents see them.

  3. 3

    Agents receive rules

    Claude, Codex, Cursor, and MCP clients receive the approved rule before they touch similar code.

Start small

Try it on one repo before changing your workflow.

Import real review history, approve the obvious wins, and expand once the rules keep proving useful.

Imports from

Rules show up where your AI coding assistant already runs.

  • GitHub PR reviews

Agents

  • Claude Code
  • OpenAI Codex
  • Cursor
  • Windsurf
  • Zed
  • Goose

Launch path

  1. Start with the free local CLI
  2. Import one repo of GitHub review history
  3. Enable Cloud approvals once one repo is working
Install diffloreView pricing

Common questions

Accepted PR review feedback, the rule your team approved, and the evidence that explains where it came from.

No. New candidates wait for approval, so noisy or one-off comments can be dismissed before agents ever see them.

Claude Code, Codex CLI, Cursor, Windsurf, Zed, Goose, and other tools that can read MCP or HTTPS context.

Yes. You can begin with the CLI and GitHub import, then use the cloud dashboard when your team wants shared review and governance.

Need a security or enterprise answer? hello@difflore.dev

Give your agent the conventions it was never trained on.

Start free: import one repo and agents get the context before the next edit.

Install difflore