
Remember decisions across agent sessions. Local, source-linked memory with CLI + MCP.
Remember decisions across agent sessions. Local, source-linked memory with CLI + MCP.
·
Stop explaining your project from scratch.
One local memory vault for your AI agents. Save the decisions, preferences and sources that matter. Ask for them again in a new session. When a decision changes, keep the correction and the history behind it.
Source alpha · Python 3.11+ · Apache-2.0
See a real memory correction · Native client test report
Give your agent this message in the project where you want memory:
Install https://github.com/Grit-77/remember-dammit by following docs/AGENT-SETUP.md. Preserve my settings. Help me save one useful fact and actually read it back. Then help me get started with the remember-dammit skill and ask whether I want no optional advisor, Jev, or Laya.
The agent setup guide covers choosing a vault and scope, previewing workspace changes, applying the configuration and completing the client's native approvals. The agent helps with installation; the package does not automatically install Python or change global client settings.
Make the first save useful. Give your agent a real project decision, a working preference or a small goal. The packaged skill asks short questions, saves selected answers as source notes and can create a task linked to your goal. It must read the saved records back before reporting success.
Then open a new session connected to the same vault and scope. Ask it to use Remember, Dammit. to recall that fact without supplying the answer again. This is explicit recall; automatic session-start memory loading is not installed.
| When this happens | Use the vault to… |
|---|---|
| “We already decided this.” | Recall the decision and the source behind it. |
| “Actually, we changed our minds.” | Correct it while retaining the previous version. |
| “I'm picking this up on another machine.” | Exchange memory between copies of the same vault. |
| “That's from a different project.” | Keep project scopes separate; grant personal notes explicitly. |
| “Where did that answer come from?” | Inspect the saved source, revision and correction history. |
Notes are readable Markdown with immutable revisions. Your client uses local CLI/MCP tools to access them. The core needs no model API key, hosted memory service, Jev or Laya; your AI client continues to use its normal account.
The demo saves an email-provider decision, corrects it, reopens the vault and finds the current decision using the former provider's name. It also checks that another project cannot recall the record. It uses synthetic data in a new directory and refuses an existing destination.
git clone https://github.com/Grit-77/remember-dammit.git
cd remember-dammit
python3 -m venv .venv
.venv/bin/python -m pip install .
.venv/bin/python examples/demo.py --vault .local/demo-vaultgit clone https://github.com/Grit-77/remember-dammit.git
cd remember-dammit
python -m venv .venv
.venv/Scripts/python.exe -m pip install .
.venv/Scripts/python.exe examples/demo.py --vault .local/demo-vaultThe command prints the actual JSON result. A captured run
recalled Use Postmark for Atlas email. using the old query SendGrid email,
retained history_states: ["superseded", "active"], and returned
other_project_records: 0.
Use a different new directory for another run. Full installation
adds the optional MCP dependency for an AI client.
| Client | Current evidence |
|---|---|
| Codex CLI | Native save/recall tested on Windows, macOS and Linux in a private synthetic run. |
| Claude Code | Workspace setup and SDK protocol checks; native trials blocked before tool use. |
| Cursor / Antigravity | Configuration and SDK checks; native client acceptance remains open. |
| Other MCP/CLI clients | The interfaces are available; compatibility needs its own verification. |
Native test report and limits · Current status
Initialize one vault, then create replicas of that identity. Independent offline contributions merge; competing corrections remain visible until an explicit resolution includes every observed head. Old copies cannot silently cancel a known retraction or access revocation. Locks and indexes stay machine-local.
Use reviewed transfer bundles or your own file transport. A private three-host Syncthing trial is documented in the native test report. The package does not install a background sync service. Raw sync is not atomic, and it cannot detect a completely unknown branch that never arrived. Existing v1 vaults need the guide's previewed, backed-up migration for these features.
This is an alpha. Capture and recall are deliberate actions. Automatic hooks, editor text merge and embeddings are not included; large-vault performance is not benchmarked. Canonical revision files are immutable—use the supported working-note workflow to edit. Implementation details and evidence
Share your first-run experience: could your agent save a useful fact and recall it in a fresh session? Where did you get stuck? Keep private notes and credentials out of the report.
If it earns a place in your workflow, consider starring the repository.
This repo does not document an install step we could read. Check the README on GitHub.
Paste this prompt into your agent. It reads this page and does the setup for you.
Read https://aiagentslisting.com/mcp/remember-dammit to learn what the "Remember, Dammit." MCP server does and how to install it. Add it to my coding agent's MCP configuration as documented on that page, then confirm the server connects and list the tools it exposes.Agents can also browse this directory over MCP at https://aiagentslisting.com/api/mcp
This server doesn't publish a machine-readable tool manifest, and it doesn't expose a public endpoint we could ask. Connect the server locally to see what it exposes.
We could not run this server: it ships no npx package for our sandbox and no public endpoint. The four checks that need a live server are left out, and the grade tops out at B until one exists.
Maintain this server? Claiming proves you own the listing and earns the last 4 points.
Claim this listingPaste this line near the top of your repository's README. The badge always renders the current score.
[](https://aiagentslisting.com/mcp/remember-dammit)Nothing comparable is listed yet.
v0.2.0a1v0.2.0a1 — first alpha packagesRelease notesConnect to the AI Agents Listing MCP server at https://aiagentslisting.com/api/mcp using the HTTP transport and complete the sign-in when it asks. Then call the submit_review tool for the listing "remember-dammit" with a rating from 1 to 5 and a short note on what worked and what did not.No reviews yet
Be the first to share your experience with this tool.
One email a week. New agents, MCP servers and skills, and what is actually getting traction.