claude-prmem 1.0.0

package/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2024 Daniel Gatis
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

package/README.md

@@ -0,0 +1,143 @@
+<div align="center">
+  <br />
+  <br />
+  <br />
+  <br />
+  <img src="logo.png" alt="prmem" />
+  <br />
+  <br />
+  <p>
+    Semantic memory for Claude Code - indexes merged PRs with AI summaries and embeddings
+  </p>
+</div>
+
+<br />
+<br />
+
+<hr />
+
+### What?
+
+Semantic memory for Claude Code. Automatically indexes your merged Pull Requests with AI-generated summaries and injects relevant context when you need it most.
+
+### Why?
+
+It's 2 a.m.
+You're fixing a bug you swear you already solved weeks ago.
+
+You ask Claude:
+"How did we handle that token refresh race condition again?"
+
+Claude has no memory.
+
+### How?
+
+1. Run `prmem sync` to index your merged PRs:
+   - Fetches PR data from GitHub (title, description, commits, files)
+   - Creates smart AI summaries of what changed and why
+   - Generates semantic embeddings for search
+
+2. All that knowledge is saved in your repo's `.prmem/` directory.
+
+3. Claude Code automatically searches your history via MCP when it needs context.
+
+### Quick Start
+
+**Prerequisites:**
+- Node.js 18+
+- GitHub CLI installed and authenticated: `gh auth login`
+
+```bash
+npm install -g claude-prmem
+cd your-project
+prmem install
+```
+
+That's it. Claude Code now has memory of your project's PR history.
+
+### Commands
+
+| Command | Description |
+|---------|-------------|
+| `prmem install` | Set up prmem in current repo (creates `.prmem/`, configures MCP) |
+| `prmem uninstall` | Remove prmem from repo. Use `-f` to skip confirmation |
+| `prmem sync` | Index merged PRs from GitHub |
+| `prmem add <number>` | Add a specific PR to memory |
+| `prmem delete <number>` | Remove a PR from memory |
+| `prmem search <query>` | Search indexed PRs semantically |
+| `prmem status` | Show indexing stats |
+| `prmem ui` | Open web UI for memory management |
+| `prmem doctor` | Check if all dependencies are installed and configured |
+
+**Sync Options:**
+```bash
+prmem sync --since 2024-01-01    # Only PRs merged after this date
+prmem sync --limit 100           # Maximum PRs to sync (default: 50)
+```
+
+**UI Options:**
+```bash
+prmem ui --port 8080             # Custom port (default: 3333)
+```
+
+### Claude Code Integration
+
+After installing prmem, Claude Code gains access to your project's memory through MCP tools.
+
+#### Smart Search Strategy
+
+When you ask about past work, Claude automatically follows this strategy:
+
+1. **Search indexed memory** using `prmem_search`
+2. **Evaluate results** - if similarity >= 80%, use them
+3. **Fallback to GitHub** - if no good results, search GitHub with `prmem_gh_search`
+4. **Index and retry** - add relevant unindexed PRs, then search again
+
+#### Available MCP Tools
+
+| Tool | Description |
+|------|-------------|
+| `prmem_search` | Semantic search of indexed PRs (returns similarity scores 0-100%) |
+| `prmem_gh_search` | Search GitHub for merged PRs (fallback when indexed results are insufficient) |
+| `prmem_status` | Check memory state: PR count, commits, files, last sync |
+| `prmem_sync` | Sync recent PRs into memory |
+| `prmem_add` | Index a specific PR by number |
+| `prmem_delete` | Remove a PR from memory |
+| `prmem_pr_commits` | Get commits for an indexed PR (paginated) |
+| `prmem_pr_files` | Get changed files for an indexed PR (paginated) |
+| `prmem_doctor` | Check if dependencies are installed and configured correctly |
+
+
+### Web UI
+
+Start the web interface to manage your indexed PRs:
+
+```bash
+prmem ui
+```
+
+### How It Works
+
+**Summaries**
+Uses your Claude Code CLI to generate intelligent summaries of each PR. No API key needed - it uses your existing Claude Code authentication.
+
+**Embeddings**
+Runs HuggingFace's `multilingual-e5-base` model locally for semantic embeddings. Downloads automatically on first sync (~400MB).
+
+**Storage**
+SQLite database stored in `.prmem/prmem.db`. Includes:
+- PR metadata (title, link, merge date)
+- AI-generated summaries
+- Semantic embeddings
+- Commit history
+- Changed files
+
+## Buy me a coffee
+
+If you find this project useful, consider buying me a coffee!
+
+[![Buy Me A Coffee](https://www.buymeacoffee.com/assets/img/custom_images/orange_img.png)](https://www.buymeacoffee.com/danielgatis)
+
+## License
+
+MIT License - see [LICENSE](./LICENSE) file.

package/dist/cli.d.ts

@@ -0,0 +1,2 @@
+
+export {  }